Bopgen MO100A phase inverter for 6550's

I've modded a Bogen MO100a to use 6550's for outputs. I got the advice below from dcgillespe. Will the attached schematic do the job for the phase inverter? I want all the beef the 6550's can stand.

6550s would also work just fine in this unit -- the high plate voltage is nothing they could not easily handle in stride, but the phase inverter stage would need a serious injection of steroids to develop the significant low impedance drive these tubes require. At a minimum, it would require powering it from a new (~ 500 volt) B+ point developed off of the output stage plate B+ source, and possibly splitting the 7247 into two separate tubes to really get the job done. A 6AV6 could be used for the input section without any changes, and 1/2 or even paralleled 12BH7 sections could handle the inverter service, with plenty of low distortion reserve drive capability for PPP pentode connected 6550 tubes.

Thank you all for your support.

6550 Driver Mod

This driver stage will work just fine. It represents a significant change from the original design, but is really what is needed. The driver stage could be tweaked a little, since the grid return resistors for 6550s should really be no more than 47K on each side -- but the 12BH7 is more than up to the task even as is.

As with any redesign however, the feedback network will require a complete reworking, which is a downside if you are trying to maintain its original characteristics. This is an upside to keeping the original driver circuits, and just turning up the gas underneath it.

The driver configuration is important no doubt, but so is the feedback network. Trading a weak driver for a great one, only to change a stable feedback network for an unstable one is no gain at all.

Just more to consider in your project!

Dave

New B+ supply

First let me thank you for your continued support. I really appreciate a pro who will take time out of their day to help a noob.

I built in a new 500 volt B+ tap as you suggested. If I read your reply correctly, you are saying it's Ok to stay with the original 12DW7 at 500 volts? This would make my life a whole lot easier.
AGAIN,,,THANK YOU FOR YOUR SUPPORT

Curley

Van -- Happy to help. If you do not have the facilities or means to properly set up and stabilize a feedback network, I would recommend that at this point, you stick with the original design, but up the voltage to the original phase inverter tube as recommended.

If the inverter stage is direct coupled from the preceding stage in the original design, it may also require some adjustment to the B+ supplying the preceding stage as well, and some adjustments to the inverter circuit component values, but these are all simple, easy changes that won't require any changes to the feedback circuits.

The end result will be much like that of a Dynaco MK III, where they supplied some serious voltage to the phase inverter circuit of that design, and drove 6550s just fine as well. The down side of this move, is that it will work the inverter section much harder -- certainly still within the capabilities of the tube -- but much harder. The up side is no feedback changes.

If you will repost the original design again, it will be easy enough to recommend inverter component values and voltage levels for the simpler conversion approach to work properly.

Dave

Original

This is the original. I used your individual bias pot suggestion and also added a second tube socket. I added a 500 volt tap to the B+ supply as well. The rest is original design. It has also been re-capped.

Thanks for your contributions. This amp is going to kick butt.

The Bogen MO100a schematic

The schematic didn't upload last time.

It's very hard to see the values on that schematic, but in essence, what you will need to do is:

1. Change the inverter B+ supply to 500 vdc
2. Change the inverter plate and cathode load resistors to matched 47K 1 Watt resistors.
3. Change the inverter cathode resistor to 1.5K.
4. Change the output tube grid return resistors to matched 47K resistors.
5. Provide a heater bias of 75 vdc.
6. Adjust output tube quiescent current to 45 ma per tube.

From a driver standpoint, this approach is not optimum, but will allow the phase inverter stage to drive a pentode 6550 output stage to full power output with little distortion. The up side, is that the modifications are rather simple to accomplish, and should yield very good results. I believe it is the most logical approach for you to take at this point in your learning curve.

The alternate driver design you provided earlier is clearly a better approach with 6550 tubes, but then requires a complete redesign of the feedback and stability circuits -- where as the more simple modification I've given should get you the vast majority of what you are looking for.

Down the road, if you want to consider adding a cathode follower stage (one dual triode tube per unit) between the existing phase inverter configuration and the output stage, that would let you have your cake, and eat it too: No changes in the feedback/stability circuits would be required, and with the addition of the CF stage, the new configuration could properly drive a quad of 6550s right out of their sockets. From a planning standpoint however, this approach would require a much more substantial negative voltage supply for proper operation (about -150 vdc at 10 ma).

Good luck with your project!

Dave

Thank you so very much. This should get me home.

OK.....I follow you but for two points. Where in the world do you adjust the heater bias? Secondly, the only resistors at the grid of the output tubes are the 1k resistors tied directly to pin 5 and the 100k pots used for bias. If you can calrify which OP tube resistors should be changed to matched 47k (point 5), I shall from this day forward know what a "grid return resistor" looks like. The learning process continues. Thank you for your knowledge.

Heater bias, would be created by moving where the center tap of the heater winding of the transformer is connected. Currently, on the original schematic, it's grounded. To create 75v positive bias, you'd need to connect that center tap to a 75v DC source.

To do this, you'd have to somehow create a +75v source- one way of doing that, is analogous to the way that Bogen created the bias supply for this amp. They used a diode, a cap, and a couple of dividing resistors, to create a low-current, moderate negative voltage supply. To create positive voltage, simply duplicate the same circuit (another supply of the same component types as the bias supply, also connected to the power transformer in the same way as the bias supply), only with the diode installed in the OPPOSITE direction (to create positive, rather than negative voltage). You'd have to change the resistor and cap values (higher resistance to ground, less resistance in series) before the diode, and probably make the cap bigger, to raise the voltage to 75v (it's considerably less than that for the original bias supply- BUT- if you're going to use 6550s, you'll need comparable voltage- around -75 to -90v - for bias for the new output tubes, so it'll need to be changed as well!).

Regards,
Gordon.

Van -- At the main B+ point supplying the output tube plates, attach a 750K 1 watt resistor. To the other end, attach a 100K 1/2 watt resistor, and a .1 uF 250 volt cap. Ground the other ends of the 100K resistor, and the .1 uF cap. At the point where all three components meet, about 73 vdc will be created. Unground the CT of the heater winding, and attach it to this point. Then, the heaters will be biased at 73 volts above ground.

In the original design, there is a 22K resistor connecting each set of output tubes to the bias source. These are called the grid return resistors, because they in fact act to return (or reference) the grids to their proper source, be it ground in cathode biased amps, or a negative bias source in fixed bias amps. They need to be raised to 47K each, and they should also be matched for best results.

Dave

Big Thank You to GORDONW for you help.

Dave, those 22k resistors were removed in the original bias mod to 100k pots for each tube. I will repost your original mod to get us on the same page.

Dave,
This is where we started:

got a schematic for the Weber ?

Honestly I haven't put any effort into figuring out what to do with mine.

Schematic for the Bogen, just in case anyone else is checking in

http://makearadio.com/schematics/ima...n-mo100a-7.jpg

I believe you'd need two more coupling caps, one from the plate of V1B to R14, and another from the cathode of V1B to R18. The existing connection from grid of V4 to V5 and from V2 to V3 would be severed. R9 and R10 would also have to go. Remove all of the cathode wiring to the balance pot and put a 10 ohm resistor from each output tube to ground.

At this point you'd have a bias supply that goes nowhere, and an audio path that feeds from the phase inverter to each tube. You should be able to put in four 100k pots from grid supply to ground, with the center wiper of each pot going to it's output tube. This would provide the adjustment per tube. The only issue is whether or not there is enough current available in the bias supply to feed that. You probably will need to play with values of R24, R25 and R26. Deleting R26 entirely might be enough to get the job done.

If you're not using it, you should be able to de-wire one of the octal accessory sockets on the top of the chassis and tie each of the cathodes to a pin on that socket. Wire one pin to ground, and you've got your bias test points without having to flip the amp over.
__________________

Van -- I would love to see the original string of posts regarding my comments above. If I made the above comments (not disputing whether I did or not), they are so general in nature that I doubt they were meant to be taken as a step by step approach, but rather as more of a working concept.

I take it from the above comments that the goal at that time was to individually bias each tube. Relating to the original design then, while using four 100K pots does successfully allow for individual biasing of each tube, there is no way I would have wanted to completely dispense with the grid return resistors. Relative to the comments above, the original grid return resistors would certainly have to go -- value and connection wise -- but not completely leave the house with Elvis.

To properly implement an individual bias modification in the original design in a step by step approach then:

1. Install the 100k bias pots as you have done.
2. Installed a 22 uF at 35 volts from each wiper of these controls to ground. Observe polarity, with the + lead going to ground (NOTE: for use with 6550 tubes, the voltage of these caps should be rated at 100 volts).
3. Remove the original 22K grid return resistors that each supplied two tubes, and install a new 47K grid return resistor for EACH output tube, supplying each tube with bias voltage from the wiper of its own bias control.

The reason the unit still worked as you had it configured, was because there were no caps from each wiper to ground. But they are needed to ensure that varying each bias control does not also change the loading on the phase inverter. The caps provide a solid AC ground at each wiper, regardless of the DC bias voltage each wiper is adjusted for. Then, with the new 47K grid resistors installed for each tube, the phase inverter always "sees" a 23.5K load (very close to the original 22K load) -- regardless of how each control is set. As it is as you have it now, the new bias controls are not only controlling the bias, but are also acting as the load on the phase inverter, altering the load on the phase inverter as bias adjustments are made, upsetting the drive balance at low frequencies.

Using this approach, the value of the original coupling caps would also need to be cut in half, since the value of each grid resistor they are supplying has been doubled.

NOW THEN, relating ALL of the foregoing to changing the output tubes over to specifically 6550s -- and in a somewhat "concept" approach (i.e. exact details to be determined):

1. Each output tube grid return resistor would be raised (doubled) again, now to 100K.
2. Each coupling cap would be reduced in value by one half, again -- now to 1/4 of the original Bogen value.
3. The bias supply would need to be altered to supply a bias voltage range of about -30 volts to -40 volts at each wiper of the new bias controls.
4. The phase inverter changes already discussed would also need to be implemented as well.

I'm sorry for any confusion in any previous comments I made regarding your project. I always try to communicate in a very clear and precise manner, so as to leave little to chance. That's why I suspect that my earlier comments were only conceptual in nature.

I hope this helps!

Dave

I can't tell you how much I appreciate your continued support of this project. As I read more and more on this site I have come to the realization that you are a "big fish" in the world of tube. To have your help means a lot. I will continue with this new information. THANK YOU AGAIN.

GordonW said:

Heater bias, would be created by moving where the center tap of the heater winding of the transformer is connected. Currently, on the original schematic, it's grounded. To create 75v positive bias, you'd need to connect that center tap to a 75v DC source.

To do this, you'd have to somehow create a +75v source- one way of doing that, is analogous to the way that Bogen created the bias supply for this amp. They used a diode, a cap, and a couple of dividing resistors, to create a low-current, moderate negative voltage supply. To create positive voltage, simply duplicate the same circuit (another supply of the same component types as the bias supply, also connected to the power transformer in the same way as the bias supply), only with the diode installed in the OPPOSITE direction (to create positive, rather than negative voltage). You'd have to change the resistor and cap values (higher resistance to ground, less resistance in series) before the diode, and probably make the cap bigger, to raise the voltage to 75v (it's considerably less than that for the original bias supply- BUT- if you're going to use 6550s, you'll need comparable voltage- around -75 to -90v - for bias for the new output tubes, so it'll need to be changed as well!).

Regards,
Gordon.

Click to expand...

One neat trick for making bias supplies is to use "un-used" filament windings to run filament TXs backwards. Say if he had a 6.3 tap, he could hook it to the secondary of a 120 by 12.6 TX secondary. Depending on things like wall voltage and actual TXs used, that would give him around 60vac, which could be stretched real close to 73 VDC. Just sayin, it's a neat trick especially when you have extra filament windings on your PT.

No extra heater winding on the MO-100a though. Space is also at a bit of a premium inside one of these things. Its actually a pretty small chassis for all the business it has going on, and what space it has is largely sucked up by the rather big power and output transformers, and the four output tubes.

also, I'm following this with some further interest myself. I'm hoping to win a bid on a tubeless MO-100a to go with my essentially tubeless MO100a. If this works out, I'll be loading them up with 6550's since I can actually get them at a price I'm willing to consider.

Jump right in

This project will end up with a pair also. Dave Gillespe is running the clinic. Couldn't be in better hands. It goes without saying you've also put a great amount of learning in my toolbox. Keep those cards and letters comin' folks!!!

well, no dice on the pair. I'll come across another eventually at a price I'm willing to pay. The great thing about being in no hurry is that you can afford to be cheap about it.

Parts are on the way. Pray for me, I need all the help I can get. Thank you all for your continued support for the noob.