Bogen CHB100 / Ampeg B15... maybe.

[ioginy]

Thanks! I totally had my head turned around on that!

[ioginy]

alrighty.
I have fixed the standby miss-hap. I also wanted to see if I could lower distortion by adjusting the NFB resistor. I read up on it a bit and the good word is that a lower resistor in the nfb loop allows more negative feedback through, causing less distortion and slightly less volume... ideal! I wired up a on/off/on switch with a 15K and a 33K resistor to try it out. Oddly though, when the switch is off (cutting off the nfb loop) the distortion and noise is lowest. As soon as I engage either of the resistors (enabling the nfb loop) the distortion JUMPS and there is a definite increase in volume. I am curious if the bogen design (which is what I used as far as the nfb loops goes) is not necessarily negative feedback, but possibly positive feedback. I really don't know much about negative feedback loops so I might be just blowing wind here, but logically if a negative feedback loop lowers distortion, a positive would increase it.... Is it possible that I have created a positive feedback loop instead?

[dcgillespie]

Yep. You have in fact created a positive FB loop. It happened because you changed output transformers, and the color code versus phasing of the replacement transformer is different from the original. The good news is that the fix is easy -- simply swap the two primary leads going to the output tubes. That will swap the phase of the feedback signal so that it in fact becomes negative FB instead of positive.

That will certainly solve the largest and most immediate problem, but may not get you completely out of the woods yet. Recall early on in this thread where I advised you to keep the small caps installed that are associated with the coupling between the AF amplifier stage, and the phase inverter stage. These help maintain HF stability in the design with the application of (negative) FB, and are transformer specific. They are most often a component that is determined by trial and error in the initial build, as their value is even somewhat determined by layout of the design as well.

These are best adjusted by running a square wave through the amplifier, and adjusting these components (or other components installed for such purpose) for the most accurate waveform, as displayed on an oscilloscope connected to the loaded output of the amplifier. If no such capability exists, the best thing to do at this point, is to simply leave the original values in place for want of anything better.

In any event, you have learned a valuable lesson in that whenever the OPT in a FB amplifier is replaced with a generic replacement, always check the FB to make sure that it remains negative. The simple test is that when the FB connection is made, gain (and therefore output signal) should be reduced. If it increases, it's positive! When that happens, reverse the primary connections as indicated.

Dave

[ioginy]

hah! Your post just made me so happy! I was thinking I would get a reply in the area of "there is no such thing as positive feedback you schmuck".

A valuable lesson indeed, thanks for the info. At the end of this project I am going to print off this thread. Not only are you folks vastly helping me my specific amp, the information that is being discussed will have a lot of use in future projects!

[soundmasterg]

A form of positive feedback is used in oscillators such as for a tremelo or vibrato, and they are setup to oscillate at low frequencies like 20 Hz or something. I added NFB to my modified Bogen CHB100 with a switch, though I am still using the stock OPT and I just sized the value (off 16 ohm tap) to where it sounded good and was enough of a difference between no NFB and with NFB to where I would use it. I often play without NFB though as the amp is more touch responsive without it.

Greg

[ioginy]

I swapped the primaries on my OPT, and installed four 1ohm resistors to check current draw on the power tubes.

For some reason now, the amp puts out almost no volume.

Also, here are the very disconcerting readings I got from the power tubes.

all tubes are getting -22.5 VDC from the bias supply

7868 #1 - 60mA
7868 #2 - 60mA
7868 #3 - 30mA
7868 #4 - 7mA

These are still the original power tubes from the Bogen as well.

I also double checked my B+ supplies and with all the tubes in and idling I am getting far better readings.

Plate supply - 510V
Output Screen supply - 400V
PI and Driver supply - 325V
Input supply - 307V

They are lower than what I was going for, but I'll fiddle with the resistor values later. I still have the 4.6K resistor in place of the 1K, which I am sure is making a difference.

[ioginy]

A form of positive feedback is used in oscillators such as for a tremelo or vibrato, and they are setup to oscillate at low frequencies like 20 Hz or something. I added NFB to my modified Bogen CHB100 with a switch, though I am still using the stock OPT and I just sized the value (off 16 ohm tap) to where it sounded good and was enough of a difference between no NFB and with NFB to where I would use it. I often play without NFB though as the amp is more touch responsive without it.

Greg

I've got my NFB hooked into a on-off-on switch with different value resistors for both "on" sides to get an idea of the values effect. Sadly the amp doesn't make much sound since I swapped the primaries of the OPT, and the NFB switch currently makes no audible difference due to that.

Looking into it now though.

[dcgillespie]

Your current readings show how unbalanced your output tubes are -- either as good tubes that are not balanced, or, as a mixed bag of tubes that has some good ones, and some bad ones. You will want to pair up tubes that require the most similar current draw for a given bias voltage, and place them on the same side of the push-pull signal. Then, you can use your separate bias controls to bring the two sides into balance.

As for signal loss, the gain should definitely be reduced due to swapping the OPT primary leads, as the original connection actually added gain to the power amplifier section, while the new connection works to reduce it. However, it shouldn't completely eliminate it, so if power is produced but requires a very high gain setting to achieve it, then gain is the problem. On the other hand, if power can't be produced regardless of much gain is added, then the power stage will need to be reviewed.

Dave

[ioginy]

I was rather afraid the tubes were seriously unbalanced. Three are visually the same with a fourth that was definitely replaced as an individual tube.

I have two brand spank'n new sets of matched 6l6's that I think i am going to convert this to. I figure, more clean headroom, WAY cheaper to replace and all I have to do is alter the bias supply a bit.

The 8 pin sockets should fit nicely into where the 9 pins were, and it looks like there's plenty of room for them.

If this is a bad idea, someone stop me because it's looking pretty good right now

[dcgillespie]

Ah, now you are into a whole different set of design considerations. But first, as to your original approach, I would imagine that your existing problem is a lack of gain, since that is what converting the FB from positive to negative accomplished. Of the 5 stages that appear before the output tubes, only 3 provide any gain. With the current power amp section requiring an estimated drive level of about .9 vac, that means that the overall gain of the preamp will need to be on the order of about 100 to properly drive the design to full power. Within the preamp section, the CF stage provides a gain loss, meaning that the requirements of the active preamp gain stages must now be at least 110 or more. But the gain of one of these stages is all but lost through the tone stack, leaving the remaining stage to do all the work. Overall, the configuration is capable of the job. However, to do that:

1. Make sure your input stage is in fact employing a 12AX7 type tube. At one point, I thought you were trying a 12AU7 in that position.

2. Use cathode bypass caps in the two active gain preamplifier stages. This is a must to achieve all possible gain.

3. Use no more than about 6 db of NFB. This level of FB will only double the drive requirement of the power amplifier section (over that of no NFB at all), and typically provide more than adequate damping for production amplifier service when pentode output stages are used. With the existing component values shown, it is estimated that the current FB level is only on the order of about 2 db.

As to converting the amplifier for use with 6L6 class tubes, consider:

1. Any gain problems you have now with 7868 type tubes will only be magnified with 6L6 class tubes. The 6L6 tube has about 1/2 the transconductance of a 7868, which will basically double the drive and bias requirements of the design.

2. Therefore, the output of the bias supply will need to be at least doubled.

3. Ideally, the grid return resistors in the output stage should be reduced to 47K when using 6L6 tubes in PPP service with fixed bias -- this to reduce the chances of thermal runaway due to reverse grid current concerns. However, this will also place a significant load on the phase inverter stage. One way to deal with this is to simply bias the tubes lightly so that they each dissipate no more than about 20-22 watts each, which should eliminate most of the concern.

4. The phase inverter stage will need particular attention to be able to drive the new output tubes. Dynaco successfully drives 6550/KT88 tubes directly from a phase splitter stage in their MK III model, which have at least as much drive requirements as 6L6 class tubes do. But the B+ to this stage is significantly elevated (to about 450 vdc or so IIRC) to accomplish this task. In addition, you will also need to remove the series 22K resistors coming from this stage to apply all available drive it can muster to the new output tubes.

5. Your new OPT will be a perfect match for 6L6 class tubes. Using such a load with a capable driver stage and 475 vdc power supply for the plates and screens, 130 watts RMS can be developed into a real world load. This is approaching the theoretical maximum power output these tubes can safely and reliably deliver.

If this shows anything, it's a good example of how using the "let's try this, or let's try that" approach often ends up with less than satisfactory results -- or at least poorer results than was had in the first place. To be successful, a given design has to work together from input to output, and from power supply to power output to achieve a target goal. Your goal can be accomplished here, but in trying to turn something into something it never originally was can often mean you end up only using the original chassis! Therefore, I guess that technically still makes this a Bogen project. On the other hand, I still have my great grandfather's ax too. It's had two new heads and four new handles since it got passed down, but it's still my grandfather's ax!

Dave

[ioginy]

1. Make sure your input stage is in fact employing a 12AX7 type tube. At one point, I thought you were trying a 12AU7 in that position.

The input tubes I am currently using are a 12AX7 first, then a 12AT7 for the second half of the cathode follower and the PI driver. I figured when testing I should keep the tube compliment as close to what I intended to use as possible.

-----

The strange thing about the "no volume" is that it sounds exactly the same when the NFB is disconnected. One of the fellows at ax84 suggested going down the line to see where the signal degrades, but it all changed when I swapped the primaries. ugh!

Either way, I think the 6L6 route is the way to go right now. It will give me the power I was originally after and I have all the parts. With the 7868's so unbalanced, it's almost a waste of time testing it, as it will never sound as it should.

1. Any gain problems you have now with 7868 type tubes will only be magnified with 6L6 class tubes. The 6L6 tube has about 1/2 the transconductance of a 7868, which will basically double the drive and bias requirements of the design.

Soundmaster at ax84 said that as far as the schematic goes, I have almost too much gain.
http://www.ax84.com/bbs/dm.php?id=446073
Increasing the bias won't be a problem at all. I designed it so I could easily swap out the first dropping resistor, as I didn't know which I would need to use in the end. YAY forethought!

I'll do a little sketching and thinking and see what I come up with. Thanks for the input!

[dcgillespie]

Greg may very well be right about the gain issue. I was just roughing some quick numbers in my head, and, it will all depend on how much loss there is through the tone stack and what your input bias is set at as well. I really can't see all of the values since the print is so small, which may be adding to the issue as well.

The input resistor really does not need to be that large as Greg said, although its contribution to noise will be insignificant as either the shorting jacks or instrument pickup will cause a much lower shunt resistance to be in place across it at all times.

Swapping the two primary leads should not have resulted in any loss of power unless something else is wrong. I agree that you should leave the FB disconnected until you find where the signal is being choked.

Dave

[ioginy]

I found a schematic for the Bogen CHB50, which uses 6l6's.
http://makearadio.com/schematics/images ... hb50-6.jpg
The phase inverter is almost exactly the same, and all in all there are very few differences.

Using this as a template I made the swap today from the 7868's to 6L6's.

The new bias supply works wonderfully, though I have a bit of a curious anomaly. I have two matched sets of 6L6's, which I put on opposing sides to get the closest current draw using the two bias taps. The strange thing is that matter what tube I put in one of the spots, it pulls almost 10mA more than all the others. The other three can be bang on 35mA, and this one will be 45mA. It's very strange.

None the less, here is the new schematic.

[dcgillespie]

For one socket to always pull more current than its mate, even when the tubes are swapped means that either: (1) something must be wired different between the two sockets, or (2) the cathode sensing resistor is off enough on the one tube reading high that its not really high at all, but just indicating high.

For 6L6s, using the plate voltage and load you are using, you will want the screens to operate at 450 vdc, and bias the tubes to 45 ma each to keep crossover distortion at bay. Tube life will still be very good at this current level, as the plate dissipation will only be at about 70% of maximum rating.

The phase inverter circuit will work as long as the required grid bias voltage does not get too high (say > -50 vdc), and you don't lower the grid return resistors from their current 100K value. Just watch your tubes to make sure the bias remains stable once set to make sure reverse grid current is not in play. I would still elevate the B+ to the phase inverter stage to ~ 400 vdc, to maximize the peak voltage this stage can deliver, and minimize its distortion as well.

Did you ever find the issue as to why the sound level was so low before?

Dave

[ioginy]

Thanks for the info. I haven't checked my b+ voltages yet, but i'll be able to get to it tomorrow.

I think I must have just had a tine stray wire touching something when I swapped the primaries of the opt. Once I tore it all down and put in the 6L6's the volume is back with a vengeance.