Suggested Dynaco PAS mods

the thermionic watercooler

Postby jukingeo » Tue Feb 10, 2009 11:52 am

Slartibartfast wrote:
As for the batteries, they would be charged while the amp is powered up.You would also start with a fresh charged set of batteries.


Oh! I didn't know that it would charge while in circuit. So then the battery would never 'die' and need to be recharged. That was my main concern.

EWBrown says I could also use a LED too. In fact I think he responded to my last post...

EWBrown wrote:The LED connects between the cathode and ground, just like a resistor (or battery). The cathode goes to ground, the anode goes to the tube's cathode, and it does not need to be bypassed, as the LED has a low effective impedance.


So basically any pre-amp tube that has about 1k for a cathode resistor (with cap) could have a LED instead? Does the LED light? If so that would be cool because the LED could double as an indicator that the stage is working.


any common 3mm low brightness red LED will work just fine.
Both halves of the 6SN7 run around 3.5 mA, the plate of the VA connects directly to the CF's grid. THe signal is than taken off the CF cathode, and run through a DC blocking cap, with a 100K to 470K resistor across the output to bleed off any DC on the output cap.


So because of the low current running through the circuit, you don't need to use any kind of resistors in series with the LED?


This will work with any common medium-mu twin triode, like 6SN7, 12AU7, 5963, 6189, ECC82, etc.


Thus it will not work with a stage that uses a 12AX7?


The schematic is similar to this one on Gary Kaufman's site:

http://www.the-planet.org/linestage2.html


Oh yeah! I seen this schematic before. As I said I have a bunch of XSN7 and XSL7 (x=insert voltage type) tubes. Most are 6volts, but I have also the 8 volt and 12 volt variants. So I must have come across this around the time I built the Rosenblit SRPP pre-amp.


Just replace the 1K VA cathode resistor with a red LED, and use a 6SN7 or 12AU7 instead of the 407A. B+ 150 to 160VDC, @ about 7 mA per channel. This may seem like a low voltage (the plate voltage is about 75V) but it operates well and sounds very good. I used a smaller output coupling cap, 1.0 uF, 400V (one of the green russian PIO caps)


Yeah, that voltage IS low. You sure that would be enough to push an amp? I typically see voltages around 225 to 300 volts on pre-amp tubes.

I used oneof the 100K stereo ladder step attenuators from Hong Kong (on e-bay) and dispensed with the 0.47 uF input cap. THe step attenuator makes a perfectly good grid resistor, and doesn't "pop" when rotated, as it is a make-before-break switch.


I am certainly going to use a 40% loudness tap control because I like the way it worked on the Rosenblit amp. (on that amp it was my ONLY tone control.

DC filaments are nice, but not truly necessary, and the filaments should be "biased" at around 37-40V above ground (or 1/4 of the B+ voltage), so that the reference voltage sits midway between the two cathode's voltages of 1.6V and 75V.
That helps to reduce / eliminate any hum or instability problems.
(This applies to SRPP drivers, as well).


Yes, I remember I had to do something like this with the Rosenblit pre-amp. It was a recommendation. However, the preamp does also have DC filaments. So it is dead quiet.

But I did know that if I do go the SRPP route with the PAS 'rebuild', I for sure would need a different power supply. SRPP's are usually picky about the type/size of power supply.


The B+ can be series or shunt regulated, if desired, but that isn't really necessary. Just icing on the cake ... Yellow_Light_Colorz_PDT_03 Yellow_Light_Colorz_PDT_02 :o


I would go that route on a custom pre-amp where I would have more room. Besides I usually like to make VR tubes visible from the outside of the pre-amp. Love the glow on those things!


Thanx for the info.

Geo
Last edited by jukingeo on Tue Feb 10, 2009 12:15 pm, edited 1 time in total.
jukingeo
 
Posts: 227
Joined: Wed Jan 21, 2009 11:45 am

Postby EWBrown » Tue Feb 10, 2009 12:05 pm

Glow is always good, except when it is happening to the power tubes' plates :o Yellow_Light_Colorz_PDT_08 And with some vintage RF power tubes, a cherry-red glowing plate is a good thing. Yellow_Light_Colorz_PDT_11

The NiCD or NiMH cells will gently trickle charge while the circuit is running, and will slowly deplete while the unit is turned off. It is good to pre-charge the cells first, to give them a better head-start.

If the 12AX7 can run with a cathode voltage of 1.25 to 1.35VDC, then the NiMH or NiCD cells can be used. I haven't yet tried them, but 880 nM Infrared LEDs operate at lower voltages, typically around 1.0 to 1.2 VDC, so they may be useful when 1.6 to 2.0V is too high.

I have run the VA section of a 6EM7 with a NiMH cell in the cathode, and the VA of a 6EM7 is pretty close to 6SL7 / 5751, and given the right conditions, this should work with a 12AX7.

refer to this design from Gary Kaufman:

http://www.the-planet.org/6EM7.html

I "breadboarded" a single channel, and it worked well and sounded quite good! I didn't go any further to build up a permanent stereo version, but I will, someday... At the time I just wanted to "prove it out" before committing it to metal.... I used a solder tab 1100 mA/H 2/3A cell just because that was what I had freely available at the time.

Actually this technique goes back to the earliest days of tubes and radio, when grid bias was supplied by low-voltage batteries. They lasted about the normal "shelf life" in this application, as properly operating class A amplifiers draw very little, or no grid current.

/ed B
Last edited by EWBrown on Tue Feb 10, 2009 1:34 pm, edited 7 times in total.
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby jukingeo » Tue Feb 10, 2009 12:59 pm

EWBrown wrote:The NiCD or NiMH cells will gently trickle charge while the circuit is running, and will slowly deplete while the unit is turned off. It is good to pre-charge the cells first, to give them a better head-start.

Actually this technique goes back to the earliest days of tubes and radio, when grid bias was supplied by low-voltage batteries. They lasted about the normal "shelf life" in this application, as properly operating class A amplifiers draw very little, or no grid current.

/ed B


Yes, I am familiar with the old A,B,C supply radios. I had an old Atwater Kent model 40 (I believe). It needed a supply like that. However, in terms of what we have been discussing, I didn't know that the battery would "charge" as the circuit was running. Being that the case, I may as well give it a shot. It could make for a nice quiet stage, especially for the phono section.

Geo
jukingeo
 
Posts: 227
Joined: Wed Jan 21, 2009 11:45 am

Postby EWBrown » Tue Feb 10, 2009 1:13 pm

I've used the NiMH or NiCD "bias" in a couple of circuits and it works quite well.

In the "old fashioned" method, the cathodes would be connected directly to ground, and the "C" bias battery would be feeding the grids through a "grid" resistor. They wouldn't get charged, but then they lasted nearly the normal shelf life. If various tubes had different bias requirements, the usual method was to use several 1.5V batteries in series, and then tap off at -1.5, -3, -4.5. -6, etc, as required. Not using resistive voltage dividers meant that there was no wasted current draw. For power tubes, cathode resistor bias was generally the preferred method.

It will trickle charge the cells at the tube's operating plate current, so, with normal usage, the batteries should last a long time.

What the batteries accomplish, is to provide the correct bias potential difference between cathode and grid, but at a much lower impedance / resistance than a standard cathode resistor and bypass cap. LEDs do this similarly.

/ed B
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby jukingeo » Tue Feb 10, 2009 9:40 pm

EWBrown wrote:Glow is always good, except when it is happening to the power tubes' plates :o Yellow_Light_Colorz_PDT_08 And with some vintage RF power tubes, a cherry-red glowing plate is a good thing. Yellow_Light_Colorz_PDT_11


Glowing screens are not that good either. One time I had an amp that had a bad plate connection on one of the EL34 tubes. The screen lit up like a power triode...yeah..WHITE! What followed was a nice beautiful blue arc inside the tube and it was toast!

The NiCD or NiMH cells will gently trickle charge while the circuit is running, and will slowly deplete while the unit is turned off. It is good to pre-charge the cells first, to give them a better head-start.


Well, as they say you learn something new every day. I always thought that you had to end up replacing the battery every so often or else use some kind of charging circuit. Some misconception THAT was.

If the 12AX7 can run with a cathode voltage of 1.25 to 1.35VDC, then the NiMH or NiCD cells can be used. I haven't yet tried them, but 880 nM Infrared LEDs operate at lower voltages, typically around 1.0 to 1.2 VDC, so they may be useful when 1.6 to 2.0V is too high.

I have run the VA section of a 6EM7 with a NiMH cell in the cathode, and the VA of a 6EM7 is pretty close to 6SL7 / 5751, and given the right conditions, this should work with a 12AX7.


Ok, so then the same principal CAN be used on tubes with higher gain. I kind of figured that because usually pre-amp tubes hover around a 1 to 3 volt bias. Thus for 3 volts I would figure you would use two batteries (or LED's) in series, right?

refer to this design from Gary Kaufman:

http://www.the-planet.org/6EM7.html

I "breadboarded" a single channel, and it worked well and sounded quite good! I didn't go any further to build up a permanent stereo version, but I will, someday... At the time I just wanted to "prove it out" before committing it to metal.... I used a solder tab 1100 mA/H 2/3A cell just because that was what I had freely available at the time.


It doesn't say what the power output of the amplifier is. Do you know the output power? I also like to breadboard things out. Surprisingly the Rosenblit amp I just built directly into the chassis.

The SE EL34 amp I am going to work on is a bit more complex and I AM going to breadboard it first.

EWBrown wrote:I've used the NiMH or NiCD "bias" in a couple of circuits and it works quite well.

In the "old fashioned" method, the cathodes would be connected directly to ground, and the "C" bias battery would be feeding the grids through a "grid" resistor. They wouldn't get charged, but then they lasted nearly the normal shelf life. If various tubes had different bias requirements, the usual method was to use several 1.5V batteries in series, and then tap off at -1.5, -3, -4.5. -6, etc, as required. Not using resistive voltage dividers meant that there was no wasted current draw. For power tubes, cathode resistor bias was generally the preferred method.


Yup I remember that being one of the ways they did it. The Atwater Kent radio I had had the huge lantern batteries for the "A" supply. Usually this was 2 or 3 in parallel. The B+ battery was considerably smaller. Usually this was two 90 volt batteries in series. This battery looked like an elongated 9 volt battery. Finally there was a 1.5v D cell for the C supply. The batteries seem to wear out in that order too. The "A" batteries would die first followed by the "B" batteries and then finally the "C" batteries. After using the radio for short time and replacing the A battery twice, I opted for a power supply. I ended up using the B batteries for a short wave regenerative radio supply later on.

It will trickle charge the cells at the tube's operating plate current, so, with normal usage, the batteries should last a long time.


Also more then likely the current going through a pre-amp tube is low enough not to cause damage or in the case of a LED, it will not blow it up. Ok, I think I am getting it now.

What the batteries accomplish, is to provide the correct bias potential difference between cathode and grid, but at a much lower impedance / resistance than a standard cathode resistor and bypass cap. LEDs do this similarly.

/ed B


So if a LED is cheaper than a battery, why the preference for the battery?

I have something else I would like to show you:

Believe it or not this is a power amp made from 5687 pre-amp tubes.

http://www.geocities.com/TimesSquare/1965/pp5687.html

Note what is connected to the cathodes of the 12AX7. This is a 1N5297 diode. You need to feed -18v into it. Supposedly it is to improve linearity, but now you need a -18v supply to do so. This was the main reason I never built this amp. Yet it is an intriguing design.

Geo
jukingeo
 
Posts: 227
Joined: Wed Jan 21, 2009 11:45 am

Postby EWBrown » Wed Feb 11, 2009 6:38 am

The direct coupled 6EM7 is about 2 watts, maybe 2.5 before it starts getting noticeably distorted. With a "true" 6EM7, the voltages I measured were in agreement with those shown on the schematic. With a 6EM7/6EA7 (much more common) the voltages were slightly "off" butit still worked fine. Changing the 120K VA plate rsistor to 130-135K would get them back to "normal" readings - just connect a 10 or 15K resistor in series with the 120K, or use two 270Ks in parallel.

For "breadboarding" I used an octal relay socket, and a Heathkit tube regulated power supply, so I could try varying the plate voltage and current. THe nice thing with the relay socket, is that it uses screw terminals, and little or no soldering is necessary, so quick changes and mods are easy.

Gary is running the power section at 50 mA, with around 180V across the tube ,and at one time he did have a 6EM7 failure, but that was just a weak or flakey tube.

I'd probably run the current a bit lower, maybe around 45-47 mA max (3K total for the cathode resistor).

The battery bias technique came along well before the idea of using LEDs, and it is still popular. Mostly personal preference, and the sound could be different, I haven't tried any meaningful side-by-side comparisons.

For higher bias voltages, red LEDs could be connected in series, or perhaps a blue or 525nM "true green" LED could be used, they are typically around 3.3 to 3.6V. The older "lime green" is around 2.2V, and amber is around 1.8 to 2V.

I like that 5687 PP design, I'd often wondered about using 5687s in SET ot PP low power amps, I've got a few dozen of them looking for "gainful employment".

Another twin triode which has similar possibilities is the 6350, with an rP of 3900 ohms, gm of 4.6 mA/V and max PD of 3.85W per section, or 7.7W both plates. THese can be paralleled, which then will have rp of 1950 ohms, gm of 9.2 mA/V and 7.7W PD, which starts becoming quite useful.

I've got LOTS of these, all vintage computer "pulls" from the old MIT COmputer Lab. All test as very good to excellent, I alredy culled out the few duds or weaklings. These, and other old "computer" tubes often show up for cheap (and sometimes even free) at the MIT fea markets held every month between April and October.

Here is another variant of that theme, from the late and great Nobu Shishido, using PPP 12BH7As:

http://www.the-planet.org/Shishido.html

It works well, I have three of these boards, one assembled prototype that Gary gave me (same one as in the photo) , and two more that I built up from bare boards.

Yeah, I'm thinking "three channel amplifier" for these... Yellow_Light_Colorz_PDT_09 :o Yellow_Light_Colorz_PDT_11

Each channel consumes about 35 mA @ 250VDC. Five watts may be a tad optimistic, but it does sound quite nice. This one just begs for a design "upgrade" to 6350s, Yellow_Light_Colorz_PDT_11


/ed B (still waking up...)
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby jukingeo » Wed Feb 11, 2009 7:50 am

EWBrown wrote:
For "breadboarding" I used an octal relay socket, and a Heathkit tube regulated power supply, so I could try varying the plate voltage and current. THe nice thing with the relay socket, is that it uses screw terminals, and little or no soldering is necessary, so quick changes and mods are easy.


I was thinking about making a 'bench' breadboard using those sockets. The breadboard for the SE EL34 amp I will be working on with be designated for just that amp. Once I get the amp running and all the 'tweaking' out of the way, I will then move all the components to a chassis. I probably will do the same thing with a Darling amp I am planning.

However for the pre-amp, I am not sure what to to. The Dynaco has the PCB's and I am not sure if I can find octal replacements for them. So if worse comes to worse, I may have to make up some metal panels with octals in them.


The battery bias technique came along well before the idea of using LEDs, and it is still popular. Mostly personal preference, and the sound could be different, I haven't tried any meaningful side-by-side comparisons.


That is what I was wondering...if there was a sound difference.

For higher bias voltages, red LEDs could be connected in series, or perhaps a blue or 525nM "true green" LED could be used, they are typically around 3.3 to 3.6V. The older "lime green" is around 2.2V, and amber is around 1.8 to 2V.


So you pretty much can use any combinations of LEDs to get the needed voltage. Cool!

I like that 5687 PP design, I'd often wondered about using 5687s in SET ot PP low power amps, I've got a few dozen of them looking for "gainful employment".


Overall, I like the design too. It was the current diode and the -18 supply requirement that stopped me from building it. I more then likely would need a power transformer that has a bias tap and I don't have any like that laying around in my arsenal (except for ridiculously huge guitar amp transformers).

Another twin triode which has similar possibilities is the 6350, with an rP of 3900 ohms, gm of 4.6 mA/V and max PD of 3.85W per section, or 7.7W both plates. THese can be paralleled, which then will have rp of 1950 ohms, gm of 9.2 mA/V and 7.7W PD, which starts becoming quite useful.

I've got LOTS of these, all vintage computer "pulls" from the old MIT COmputer Lab. All test as very good to excellent, I alredy culled out the few duds or weaklings. These, and other old "computer" tubes often show up for cheap (and sometimes even free) at the MIT fea markets held every month between April and October.


I got a good deal on a BUNCH of 6080's, that nice big fat dual triode bottle used in voltage regulators. The tube is a bit 'outdated' for it's original application and since I got a bunch of them, I was thinking about an OTL design. But I learned of the problems with those amps and was wondering what other amp (transformer based) designs I could build with that tube. If I can't find anything good, I am probably going to sell the tubes off. I have seen a cool headphone amp based on the 6080, but that is about it.


Here is another variant of that theme, from the late and great Nobu Shishido, using PPP 12BH7As:

http://www.the-planet.org/Shishido.html

It works well, I have three of these boards, one assembled prototype that Gary gave me (same one as in the photo) , and two more that I built up from bare boards.

Yeah, I'm thinking "three channel amplifier" for these... Yellow_Light_Colorz_PDT_09 :o Yellow_Light_Colorz_PDT_11

Each channel consumes about 35 mA @ 250VDC. Five watts may be a tad optimistic, but it does sound quite nice. This one just begs for a design "upgrade" to 6350s, Yellow_Light_Colorz_PDT_11


/ed B (still waking up...)


12BH7??? Isn't that a fairly expensive tube? I remember using these in the Ampeg SVT. I don't remember them being cheap. Then again, nothing is cheap when it comes to an SVT. In case you may not know, an SVT is an all tube bass guitar amplifier that can put out 300 watts into a 4 or 2 ohm load. It uses 6 6550's in the output section. It is a VERY mean amplifier and a force to be reckoned with.

Yeah...I am still waking up too. ...coffee isn't working this morning. I have to work on tube amps today too. Certainly wouldn't want to brush against pin 3 or 4 on an output tube and wake up "the hard way".

Geo
jukingeo
 
Posts: 227
Joined: Wed Jan 21, 2009 11:45 am

Postby EWBrown » Wed Feb 11, 2009 10:36 am

12BH7As can be pricey, but I havve a modest selection of NOS and good used ones to pick from. WIth a filament re-wire, 6GU7s can be used, and I'd wager that 6350s can slip right in there, maybe have to twiddle the cathode resistor a bit for peak performance.

Looking at the 5687 SE and PP designs on Bob Danilek's site, it becomes apparent that a paralleled 5687 could be a very good stand-in for a 1626, though the cathode resistor should be half the value for a 1626 (500 ohms inseead of 1000) . The "numbers" otherwise work out fairly closely Yellow_Light_Colorz_PDT_11

I was looking at the current diodes at Mouser. The 1 mA and 4.7 mA are about $5 a pop. However, 1.1 mA and 4.3 mA are less than half that cost.
I'll order some 1.1, 2.2, 3.6 and 4.3 mA current diodes to play around with. These should come in handy for most tubular applications.

The 4.7 mA current diode (1N5314) is used in a 6SN7 gain stage in his 2A3 PPP amp. Here:

http://www.geocities.com/bobdanielak/project.html


I suppose that 4.3 mA should also wprk just fine.

Then...

There is no good reason that LM334Z couldn't be used for these applications, and they are easily adjustable with a single resistor.

The -18V supply could be generated with a small 12V tranny, wall wart, or siphoned off the main power trannie secondary with a capacitive and resistive "tap". The current is only 1 mA (or 1.1 mA) so there isn't much current demand there.

Actually you could use -9V (or less) as the main reason of having the 1N5297 referenced to a negative voltage is to keep it well above its "pinch off" voltage of around 1.35V. So anything from -6 to -18 should work. The 4.3 mA with a 6SN7, 12AU7 or 12AT7 is operating beyond the "pinch off" point so they can go to the ground buss.

That 5687 PP amp's 12AX7 driver / PI circuit could be useful with EL84s, 6S4s, 6350s and other output tubes with a 10-14 V grid drive. Or it could be applied to 6EM7 VA stages with a little "tweeking". I'm still trying to "perfect" the 6EM7 PP design, thee original PI works decently, but it could be better.

Going back to the PAS preamps (we have drifted just a bit off course). the operational equivalents of 6SL7 and 6SN7 are the 5751 and 6CG7.

Trying to fit the octal tubes into a PAS chassis, and also keep the cover on, may be a real exercise in mechanical engineering wizardry. :o

I suppose, that with some internal brackets, the PC boards could be used, then wire some octal sockets to the boards, and then mount the sockets and tubes horizontally. The "long axis" of the plates ideally should be in a vertical plane, for horizontal mounting.

Sometimes I do like the 40% tap "loudness control" capability. I just have to "translate" the proper external R and C values for the 100K Rat Shack "alps" stereo pot. Just another "get around to it" project for me to complete someday... I've "Swagged" it by dividing fromthe original SCA-35 volume / loudness controls. but it just didn't sound "right"...

This could also be applied to series step attenuators, but not so easily for the ladder type step att's.

Bob Danilek has a 6SL7 phono preamp here:

http://www.geocities.com/bobdanielak/dc_phono.gif

and 12SL7 / 12SN7 here:

http://www.geocities.com/bobdanielak/phono.html

/ed B
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby jukingeo » Wed Feb 11, 2009 1:43 pm

EWBrown wrote:12BH7As can be pricey, but I havve a modest selection of NOS and good used ones to pick from.


Oh, that is like in my case with the 6SN7's. I have TONs of them. I saw a box full of them (most NOS) in a TV shop and asked if the guy still used them, he said, "Nah", and I bought the whole lot from him :). Then there was the old TV/Radio shop that closed down and that is where I acquired some of odd voltage variants (such as the 8SN7's, 12SN7s, 12SL7's.

Looking at the 5687 SE and PP designs on Bob Danilek's site, it becomes apparent that a paralleled 5687 could be a very good stand-in for a 1626, though the cathode resistor should be half the value for a 1626 (500 ohms inseead of 1000) . The "numbers" otherwise work out fairly closely Yellow_Light_Colorz_PDT_11


Really? I kept the design in my computer because I thought it was pretty cool to get around 4 to 8 watts out of that little tube.

I was looking at the current diodes at Mouser. The 1 mA and 4.7 mA are about $5 a pop. However, 1.1 mA and 4.3 mA are less than half that cost.
I'll order some 1.1, 2.2, 3.6 and 4.3 mA current diodes to play around with. These should come in handy for most tubular applications.


I have not played around with current sourced designs, but I know that I have a few schematics in my archives that do show current references on the pre-amp tubes. As I said, I mostly shied away from them because of the 'extra' power supply requirements. As it is I most certainly would try a Ni-Cad battery before 'adding MORE iron' to an already heavy amp (or pre-amp) design. The only way I would consider it is (as I said), if I can get a smaller transformer that has a bias tap. Then with the bias tap I could create the necessary supply needed for the current reference.



The 4.7 mA current diode (1N5314) is used in a 6SN7 gain stage in his 2A3 PPP amp. Here:

http://www.geocities.com/bobdanielak/project.html


I suppose that 4.3 mA should also wprk just fine.

Then...

There is no good reason that LM334Z couldn't be used for these applications, and they are easily adjustable with a single resistor.


Yeah, I seen that Bob used a couple of those diodes in that amp design. I have seen the LM334Z used as well.


The -18V supply could be generated with a small 12V tranny, wall wart, or siphoned off the main power trannie secondary with a capacitive and resistive "tap". The current is only 1 mA (or 1.1 mA) so there isn't much current demand there.

Actually you could use -9V (or less) as the main reason of having the 1N5297 referenced to a negative voltage is to keep it well above its "pinch off" voltage of around 1.35V. So anything from -6 to -18 should work. The 4.3 mA with a 6SN7, 12AU7 or 12AT7 is operating beyond the "pinch off" point so they can go to the ground buss.

That 5687 PP amp's 12AX7 driver / PI circuit could be useful with EL84s, 6S4s, 6350s and other output tubes with a 10-14 V grid drive. Or it could be applied to 6EM7 VA stages with a little "tweeking". I'm still trying to "perfect" the 6EM7 PP design, thee original PI works decently, but it could be better.


Ehhh, I don't know, I do still have my reservations about putting a piece of silicon in the audio path. So that could be one hold off on my part about using LED's and perhaps the battery method WOULD be better. But I am not sure, it does warrant a comparison, that is for sure.

Going back to the PAS preamps (we have drifted just a bit off course). the operational equivalents of 6SL7 and 6SN7 are the 5751 and 6CG7.


Yes, we have (oops! sorry!). I have heard of the 5751 before, but not the 6CG7. Anyway, as I said, I don't have to bother with using a cheaper alternative tube, since I have many of xSN7 and xSL7 variants. So I am good to go here. As far as mini-tubes go, as I said, I fix guitar amps, so I can get just about all the tubes in the 12A_7 volt series (12AX7, 12AT7, 12AU7, 12DW7, 12AH7 (that is how I know these are expensive)).

Trying to fit the octal tubes into a PAS chassis, and also keep the cover on, may be a real exercise in mechanical engineering wizardry. :o


Well, I would like to go that route if I CAN do it. But it doesn't have to be octals. After all they are buried inside the piece and not displayed outside. So, I can use novel mini tubes too. In which case I would stick with the 12AX7 for the phono stage and perhaps go with a 12AU7 for the line stage.

I suppose, that with some internal brackets, the PC boards could be used, then wire some octal sockets to the boards, and then mount the sockets and tubes horizontally. The "long axis" of the plates ideally should be in a vertical plane, for horizontal mounting.


Again, if the tube doesn't go in upright, I am not going to sweat it. I am sure down the road there will be plenty of opportunities to create an 'octal only' preamp.

Sometimes I do like the 40% tap "loudness control" capability. I just have to "translate" the proper external R and C values for the 100K Rat Shack "alps" stereo pot. Just another "get around to it" project for me to complete someday... I've "Swagged" it by dividing fromthe original SCA-35 volume / loudness controls. but it just didn't sound "right"...


I 'stole' my loudness circuit from the amp designs in the Mullard book. It sounds really nice. When I was testing the pre-amp at my former job, one of the guys was completely floored by the sound. I was using a solid state MOSFET amp though, so I had some pretty good sound out of it.


This could also be applied to series step attenuators, but not so easily for the ladder type step att's.

Bob Danilek has a 6SL7 phono preamp here:

http://www.geocities.com/bobdanielak/dc_phono.gif

and 12SL7 / 12SN7 here:

http://www.geocities.com/bobdanielak/phono.html

/ed B


This circuit here supposedly ups one on Bob's phono preamp. I was reading this the other day:

http://home.earthlink.net/~jeremyepstein/6c45riaa.html

I am not too familiar with that tube though. However if you notice the last schematic he apparently built a whole pre-amp around it. He is using a single stage for the line amp which is a higher impedance drive.

Geo
jukingeo
 
Posts: 227
Joined: Wed Jan 21, 2009 11:45 am

Postby TerrySmith » Wed Feb 11, 2009 5:39 pm

Well I thought I would give the LED bias a try. Seems to be Quieter, maybe less rumble. Still sounds very good!

Image

Image
T. Smith
User avatar
TerrySmith
KT88
 
Posts: 973
Joined: Sat Jan 31, 2004 12:51 pm
Location: Maryville TN

Postby EWBrown » Wed Feb 11, 2009 7:31 pm

The red LED technique has been around a few years. Doc Bottlehead uses it in a lot of their designs, in linestages, phono preamps (the Seduction) and in their C4S (Camille's Cascoded Constant Current Source. These work in concert with a 2N2907A, MJE350 (both PNP) and a couple of resistors, and the two HLMP6000s as voltage references.

The LED of choice is the HLMP6000, it was HP, now they go under the name of Avago. (sounds like Steve Irwin - ('ave a go at that, mate) Yellow_Light_Colorz_PDT_04

I get mine from Jameco, they are relatively cheap. The HLMP6000 won't win any brightness contests, but they are very consistent, I found they operate between 1.56 and 1.57 VDC as cathode "resistors" and most often they dial right in at 1.567V

These have thin, flat "ribbon" leads, and the cathode end is marked on teh bottom with a silver band. They are physically hallfway between surface mount and leaded devices.

Unlike CCS devices, which operate with a very high effective impedance, LEDs operate as constant voltage sources, and have a very low effective impedance (about ten ohms) , and do not require capacitor bypassing. They seem to have less noise self-generation than most resistors.

For most generap purpose uses, any common red LED with a VF drop of 1.8 to 2.9V is fine, the HLMP6000s come in handy when 1.567 V is a better value.

This is briefly covered in the G*S*G foreum, as replacements for the 3K cathode resistors for the 6SL7., which operates with a plate current of approximately 0.6 mA. That equates to 1.8V, and a "common" red LED fits the bill perfectly.

An interesting circuit "hack" to try (I already bounced this one off Shannon), and I've used it in numerous circuits:

In place of the 3K resistors and bypass caps for the 6SL7 cathodes, use a "cheap red LED", cathode to ground, anode to the tube's cathode.
This offers the same 1.8VDC offest. A good location for the LEDs is the C3 and C4 positions, orient the LEDs so the cathode is to ground, and its anode is to the + side.

Since the current we are dealing with is only 0.6 mA, either a 2mA or 20 mA red LED can be used. My LED of choice is the HLMP1700 or HLMP4700.
HLMP6000 is also a good choice, I've used these as "cathode resistors" in other linestage and phono stage designs, with good results.

LED brightness isn't a factor here, the old 3mm or 5mm dim ones with the diffused "bullet" are good enough. The super-high-brightness retina-zappers aren't needed (or desired) here.

These LEDs can be considered as a "poor man's" constant voltage device. They do not have to be bypassed.


/ed B in NH
Last edited by EWBrown on Thu Feb 12, 2009 9:04 am, edited 2 times in total.
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby jukingeo » Wed Feb 11, 2009 8:39 pm

TerrySmith wrote:Well I thought I would give the LED bias a try. Seems to be Quieter, maybe less rumble. Still sounds very good!


Lookin' really good. Looks like a nice simple, uncluttered and straight to the point design, which is something I am always interested in. I noticed the line stage is different from the pictures you posted on page 2 of this thread. Do you have the full schematics (both phono and line stage) for this?

EWBrown wrote:
Unlike CCS devices, which operate with a very high effective impedance, LEDs operate as constant voltage sources, and have a very low effective impedance, and do not require capacitor bypassing. They seem to have less noise self-generation than most resistors.


Now the one thing I don't get though is that with a LED being a diode, wouldn't it have a rectifying effect on the audio signal which is AC? It is after all a diode. So that is something that escapes me.

But from what Terry describes (above) with his application, I am really willing to give it a go.


THis is briefly covered in the G*S*G foreum, as replacements for teh 3K cathode resistors for the 6SL7., which operates with a plate current of approximately 0.6 mA. That equates to 1.8V, and a "common" red LED fits the bill perfectly.

/ed B


Did you mean 6 ma or 0.006 A? 0.6ma is kind of low for plate current.

Geo
jukingeo
 
Posts: 227
Joined: Wed Jan 21, 2009 11:45 am

Postby nyazzip » Wed Feb 11, 2009 10:15 pm

so don't light emitting diodes produce "switching noise" too, like rectifier diodes are alleged to do?
<i>the poor craftsman blames his tools</i>
User avatar
nyazzip
KT88
 
Posts: 1073
Joined: Wed Jan 09, 2008 1:24 am

Postby jukingeo » Thu Feb 12, 2009 7:10 am

nyazzip wrote:so don't light emitting diodes produce "switching noise" too, like rectifier diodes are alleged to do?


This is what I was wondering as well. It seems as though Terry has created a successful phono stage. So if the LED's pass THAT test then it should be good.

But I am happy to see the results are positive because I am always adamant about accepting the use of silicon anywhere near the audio path in a tube amp.

So what I am thinking is that if the LED doesn't rectify the audio then it has to ride on top of the DC. What would be cool is if Terry has an oscilloscope if he could tap directly into the cathode and push a sine wave into the stage to see what is happening. Perhaps he could also test for any noise the stage could be generating using the diode in comparison to using a standard resistor/capacitor bypass. Then that would put it to rest for sure.

Whadja think Terry? Got an audio generator and an 'O'-scope?

Geo
jukingeo
 
Posts: 227
Joined: Wed Jan 21, 2009 11:45 am

Postby EWBrown » Thu Feb 12, 2009 8:52 am

In the GSG, the 6SL7 runs at 0.6 mA (600 microamps) plate current.

That works for high gain VA stages.

The 12AX7 phase splitter / driver inteh 5687 PP amp runs at 1 mA for both sections, that's 500 microamps per triode.

I've not had any noise issues with LEDs in the cathode circuit, even in a high-gain phono stage using 6DJ8s and CCS plate loads (Bottlehead Seduction). LEDs won't produce any "switching noise" like rectifier diodes, because in this application, they are handling a relatively constant DC current and voltage.

I know, it flies in the face of all "old school" tube edsign experience, and I resisted even thinking about trying it for years. But I got curious and gave it a try, and was pleasantly surprised.

As the "BORG" say, resistance is futile, you will be assimilated Yellow_Light_Colorz_PDT_08 Yellow_Light_Colorz_PDT_03 :o

/ed B
Last edited by EWBrown on Thu Feb 12, 2009 9:15 am, edited 3 times in total.
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

PreviousNext

Return to diy hifi

Who is online

Users browsing this forum: No registered users and 9 guests

cron