Help understanding some things about amplifier design

ask your general tube related questions here

Help understanding some things about amplifier design

Postby mrtubenewbie » Tue Aug 02, 2016 12:08 pm

Hi all,

Until recently I'd always seen vacuum tubes as mystical things that I could never understand. A few weeks ago I had some time to myself and decided that I was going to learn how they work and build a small amplifier. I've been reading a book called 'Vacuum Tube Amplifier Basics' which has given me a lot of information, a lot of which I think I get, but there are still parts of it that haven't quite sunk in... I was hoping that I could outline my plan so far and maybe someone with experience can highlight the mistakes I'm making or help me to understand some things. I'm very new to this, so I apologise in advance for any massive errors I make (???)

So I guess I'll state my goal first - as this is primarily an educational / fun project, I'm not looking for something perfect - just something that I can comfortably use.
Currently I'm thinking of a class AB amplifier using a pair of 6V6GTs and a 12AX7 per channel. For the power supply I was thinking about one of the Edcor power transformers like the XPWR010 that will output 750V @ 175mA (with a 120 or 240V input) and a 5U4G/5U4GB rectifier, connected to a choke filter. I've attached a design (pretty much straight from the book I mentioned) using a 1n5942 zener for about 50V max bias. I could probably switch to a different rated zener as the 6V6GTs only appear to want a -15 to -19V grid voltage based on the configurations listed in the datasheet. This is pretty much the first thing I've ever tried drawing up on KiCad so I'm sorry for how bad it is

...so anyway if I skip to the tubes - the tung-sol datasheet for the 6V6GT lists two configurations for a class AB amplifier. The second seems nicer - listing about 14W output through an 8000 ohm load. It does say that for 2 tubes the max. signal plate current would be about 92mA though - which with 2 channels would surely be over the 175mA maximum of the transformer mentioned above, so I guess I'll have to pick a new transformer or operate the tubes closer to the first listed configuration, which has an output closer to 10W through a 10,000 ohm load and a max. signal plate current of 79mA for 2 tubes. I'm probably misunderstanding something here...

I'd like to be able to drive the amplifier from a (consumer) line-level signal, so I guess that would be 0.4v-ish peak input? The 6V6GT datasheet shows (again in the typical configurations) either a 30 or a 38V peak AF input voltage. If I'm understanding the book I've been reading the gain I need is 30/0.4 := 75 or 38/0.4 := 95. Also from the book, apparently a rough estimate of a tube's max voltage amplification is (Mu * Rp) / (Rp + Ra), so I guess with a 100k plate resistor a 12AX7 would give me (100 * 100,000) / (100,000 + 62,500) := 61.538 (assuming the typical configuration of plate voltage: 250V in the GE datasheet for the 12AX7). I guess I'd need to run the signal through both of the 12AX7's triodes in order to amplify the voltage enough to fully drive the 6V6GTs.
[side note: I haven't quite understood how to choose the plate resistor, I've read around that it should be about 3 * the internal plate resistance, but I don't yet have a solid understanding on how it affects amplification, or how to pick a second plate resistor for the other side of a 12AX7 - should it just be the same value?]

Is it a bad idea to run a 12AX7 this way? I'm having trouble figuring out how to put the two sides together, I've seen a bunch of different schematics on the Internet, some with a cathode bypass capacitor on the first triode, some on the second, some neither. Some with a larger plate resistor on the first and a smaller (or no resistor?) on the second, or the reverse... it all seems a little conflicting. Then comes the phase splitting, as I want to run the power output tubes in a class AB setup, would it be advisable to introduce another tube that does not do any voltage amplification just to split the output?

Another bit of confusion I have is all the way back at the power supply. So I'm in the UK and mains voltage is 230V, so if the transformer I mentioned is outputting 750V @ 175mA when the primary is at 240V, wouldn't the output at 230V when fully loaded be closer to (750/240)*230 := 718.750? Then if the bias voltage is subtracted from that (and still assuming a 51V zener), wouldn't I be feeding the 5U4GB (718.750 - 51) / 2 := 333.875 per plate?
I've had trouble understanding the output voltage graph for the 5U4GB with a choke filter - the voltage drop curve - is that for a constant input voltage? Because unloaded (or with reduced load) wouldn't the transformer's output voltage be a bit higher? Assuming even the 350V input/plate line, by 175mA-ish it looks as though the output voltage would be closer to 260V or so... Is my current "design" running too low voltage?

Sorry for the wall of text, there's a lot going on in my head regarding this amplifier right now, and I'm feeling a little confused in multiple ways... I don't want to just go out and buy some kind of kit, because I want to actually understand this stuff a little better and actually (at least in part) design something that I can then build. I guess that will take time though. :(
Attachments
whatamidoing-psu.png
whatamidoing-psu.png (11.73 KiB) Viewed 5841 times
mrtubenewbie
 
Posts: 2
Joined: Mon Aug 01, 2016 3:09 pm

Re: Help understanding some things about amplifier design

Postby mike567 » Tue Aug 02, 2016 7:32 pm

Download psud2 from duncan amps. http://www.duncanamps.com/psud2/index.html
mike567
 
Posts: 24
Joined: Sun Jan 08, 2012 10:14 am

Re: Help understanding some things about amplifier design

Postby Geek » Tue Aug 02, 2016 10:37 pm

Might want to rethink that bias control as well. That configuration is known as "Suicide Bias", because when the pot ages and the wiper lifts, you have dead output tubes and possibly an open OPT winding.

Cheers!
-= Gregg =-
Fine wine comes in glass bottles, not plastic sacks. Therefore the finer electrons are also found in glass bottles.
User avatar
Geek
KT88
 
Posts: 3585
Joined: Sun Oct 21, 2007 3:01 am
Location: Chilliwack, British Columbia

Re: Help understanding some things about amplifier design

Postby mrtubenewbie » Wed Aug 03, 2016 12:04 pm

Geek wrote:Might want to rethink that bias control as well. That configuration is known as "Suicide Bias", because when the pot ages and the wiper lifts, you have dead output tubes and possibly an open OPT winding.

Cheers!


Ah okay, should I just decide on my bias voltage and lose the pot? I guess as the 6v6gt only wants about -15 to -19v I could use a 20v or so zener...

mike567 wrote:Download psud2 from duncan amps. http://www.duncanamps.com/psud2/index.html

Giving this one a go now. I used the off-load voltage calculator and set it to 349.375 @ 175A (what I thought the output would be with 230V on the primary, with a 20v bias zener). With a constant-current load (which I guess isn't really that realistic) of 150mA-188mA and the voltage at the load seems to be around 269-275v -ish. Do you think it's safe to design the rest for a B+ voltage within this range?

I realise I'll also need an additional tube to use as a phase inverter. Would another 12AX7 be a good choice or should I find something else?


Again, sorry for so many questions - and thank you for your help!
mrtubenewbie
 
Posts: 2
Joined: Mon Aug 01, 2016 3:09 pm

Re: Help understanding some things about amplifier design

Postby Geek » Wed Aug 03, 2016 10:58 pm

mrtubenewbie wrote:Ah okay, should I just decide on my bias voltage and lose the pot? I guess as the 6v6gt only wants about -15 to -19v I could use a 20v or so zener...


No, that would be worse. A zener will prevent the bias from following line voltage fluctuations (hence B+ variations). Also, each tube type biases differently. Cheaper brands need more than "top shelf" brands.

Here is a failsafe example where if the wiper lifts, the tube just cuts off:
http://geek.scorpiorising.ca/contrib/Ge ... y_bias.png

Cheers!
-= Gregg =-
Fine wine comes in glass bottles, not plastic sacks. Therefore the finer electrons are also found in glass bottles.
User avatar
Geek
KT88
 
Posts: 3585
Joined: Sun Oct 21, 2007 3:01 am
Location: Chilliwack, British Columbia

Re: Help understanding some things about amplifier design

Postby Impmon » Fri Dec 16, 2016 11:41 pm

You bring up a number of different issues here.

mrtubenewbie wrote:Hi all,

So I guess I'll state my goal first - as this is primarily an educational / fun project, I'm not looking for something perfect - just something that I can comfortably use.
Currently I'm thinking of a class AB amplifier using a pair of 6V6GTs and a 12AX7 per channel. For the power supply I was thinking about one of the Edcor power transformers like the XPWR010 that will output 750V @ 175mA (with a 120 or 240V input) and a 5U4G/5U4GB rectifier, connected to a choke filter. I've attached a design (pretty much straight from the book I mentioned) using a 1n5942 zener for about 50V max bias. I could probably switch to a different rated zener as the 6V6GTs only appear to want a -15 to -19V grid voltage based on the configurations listed in the datasheet. This is pretty much the first thing I've ever tried drawing up on KiCad so I'm sorry for how bad it is


A 12AX7 based LTP splitter should be able to drive 6V6s, as it's not a too difficult load.

...so anyway if I skip to the tubes - the tung-sol datasheet for the 6V6GT lists two configurations for a class AB amplifier. The second seems nicer - listing about 14W output through an 8000 ohm load. It does say that for 2 tubes the max. signal plate current would be about 92mA though - which with 2 channels would surely be over the 175mA maximum of the transformer mentioned above, so I guess I'll have to pick a new transformer or operate the tubes closer to the first listed configuration, which has an output closer to 10W through a 10,000 ohm load and a max. signal plate current of 79mA for 2 tubes. I'm probably misunderstanding something here...


You don't have to do that. The 92mA is a peak value whose average is 58.6mA, or 117.2mA for two channels. Even that applies only under "brick on the key" conditions of outputting constant max RMS power. Audio is very far from such conditions. You'll be working closer to the no-signal value of 70mA, or 140mA for both channels, so the PTX looks good here so far as the current rating.

I'd like to be able to drive the amplifier from a (consumer) line-level signal, so I guess that would be 0.4v-ish peak input? The 6V6GT datasheet shows (again in the typical configurations) either a 30 or a 38V peak AF input voltage. If I'm understanding the book I've been reading the gain I need is 30/0.4 := 75 or 38/0.4 := 95.


Output levels for CD players, tuners, etc is 1.0Vrms, so more like 1.4Vp. The open loop gain needs to be: 19/1.4= 13.57 per phase. Any extra open loop gain can be devoted to gNFB.

Also from the book, apparently a rough estimate of a tube's max voltage amplification is (Mu * Rp) / (Rp + Ra), so I guess with a 100k plate resistor a 12AX7 would give me (100 * 100,000) / (100,000 + 62,500) := 61.538 (assuming the typical configuration of plate voltage: 250V in the GE datasheet for the 12AX7). I guess I'd need to run the signal through both of the 12AX7's triodes in order to amplify the voltage enough to fully drive the 6V6GTs.
[side note: I haven't quite understood how to choose the plate resistor, I've read around that it should be about 3 * the internal plate resistance, but I don't yet have a solid understanding on how it affects amplification, or how to pick a second plate resistor for the other side of a 12AX7 - should it just be the same value?]


You can't do hollow state design by formula. This has to be done by loadline analysis. 100K is a bare minimum for 12AX7s, as triodes like light plate loads for the best linearity. 120K+ would be better.

Is it a bad idea to run a 12AX7 this way? I'm having trouble figuring out how to put the two sides together, I've seen a bunch of different schematics on the Internet, some with a cathode bypass capacitor on the first triode, some on the second, some neither. Some with a larger plate resistor on the first and a smaller (or no resistor?) on the second, or the reverse... it all seems a little conflicting. Then comes the phase splitting, as I want to run the power output tubes in a class AB setup, would it be advisable to introduce another tube that does not do any voltage amplification just to split the output?


Your best bet would be to use the 12AX7 as an LTP with active tail loading. This provides the best AC balance: both phase-to-phase and harmonic distortion. You can use that as the sole gain stage, or stick a second differential gain stage after it if you need more open loop gain.

Another bit of confusion I have is all the way back at the power supply. So I'm in the UK and mains voltage is 230V, so if the transformer I mentioned is outputting 750V @ 175mA when the primary is at 240V, wouldn't the output at 230V when fully loaded be closer to (750/240)*230 := 718.750? Then if the bias voltage is subtracted from that (and still assuming a 51V zener), wouldn't I be feeding the 5U4GB (718.750 - 51) / 2 := 333.875 per plate?
I've had trouble understanding the output voltage graph for the 5U4GB with a choke filter - the voltage drop curve - is that for a constant input voltage? Because unloaded (or with reduced load) wouldn't the transformer's output voltage be a bit higher? Assuming even the 350V input/plate line, by 175mA-ish it looks as though the output voltage would be closer to 260V or so... Is my current "design" running too low voltage?


Doing some very preliminary calcs, the DC rail voltage looks close, assuming you are going with 5U4GBs and choke input. You would be overvolting badly with a C-input filter and/or substituting SS power diodes. Even if your rail is down around 260V, it doesn't make much difference. Hollow state is a good deal more "forgiving" than solid state, and it's still within +/-10% anyway.
Live Free or Die: Linux
User avatar
Impmon
 
Posts: 68
Joined: Sat Jun 14, 2008 8:05 pm
Location: The Midwest


Return to tube 101

Who is online

Users browsing this forum: No registered users and 1 guest

cron