• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

The Mighty Bogen MO-200A

dcgillespie said:
The level control has been removed and replaced with a new Bias Control that -- once the DC Balance Control is set -- simultaneously moves both output tubes up or down in quiescent current draw, so that the DC balance is maintained while the lowest distortion operation point is determined.
Click to expand...

Ah, now I get what you were trying to tell me a year ago when I built my 60 watt amp. The part I was missing was the ability to adjust bias until the lowest distortion operating point is identified. Can be done with individual bias adjust controls I suppose, just a bit harder to reposition bias on two tubes while maintaining balance. Makes sense now.
 
It really is the only way to go when the goal is to balance the quiescent current of the tubes -- AND -- adjust the stage for the lowest distortion operating point. Trying to successfully achieve that goal with individual bias controls per tube is nearly impossible.

This approach solves the problem completely, and when implemented correctly, provides a fail safe element as well. In virtually all adjustable bias designs, if a control wiper lifts, the tubes are immediately in trouble. By using the bias/balance approach with the wipers on the ground side of the circuit, if they ever lift, the tubes simply receive too much bias voltage, and cool down to a near cut off state. It is absolutely having your cake, and eating it too!

Dave
 
Thanks Dave, this is by far the best discussion threads that I have read in a long while, lots of valuable information here. Slightly OT, but do you have some information on your bench setup - you seem to have some custom-built test gear, inquiring minds want to know ;-)
 
Hi Jaz -- Thanks for the kind words. My lab equipment has evolved over the years with my original scope being one I built as a boy (Eico), and my first distortion measuring equipment being the really excellent Heath IM-5258 and IM-5248 THD and IMD analyzers respectively. My everyday workhorse signal generator is also a Heath IG-5218.

As I expanded into more critical work, and work requiring greater bandwidth than that of the basic audio spectrum (FM and MPX work), I upgraded my bench many years ago, so that now, my go to distortion analyzer is the superb HP 339A, and my DVM is also HP -- the 3466A -- whose calibration accuracy is traceable to the NIST. My scope is a workhorse Tek 475A. I also very much enjoy my HP 400 and 3400A AC and True RMS analog AC voltmeters as well. Finally, the remainder of my primary commercial gear is a Heath IM-4180 FM Deviation Meter, and a Fisher 300 Multiplex generator. There are other pieces as well -- an RF generator, Bridge, ect.

My custom equipment is my tube testers -- a power output tester for power tubes, and a Gm/AC Gain tester for small signal tubes. Good commercial tube testers are fine for culling out the shorted, dead, and gassy tubes, but without operating the tubes at real world voltage levels, give you only a very general picture of a tube's condition.

My power output tester is designed to operate a power tube under maximum rated cathode current under dynamic conditions, and the Gm tester tests small signal tubes under conditions outlined by RCA, Sylvania, and others in their tube manuals. With these testers, I can get a very precise picture of tube health as well as operating characteristic, that allows for extremely close grading and matching of power tubes, and small signal tube sections.

Thanks for the interest!

Dave
 
Dave, how many volts of balance adjust did you allow per side on this amp? Seems like 5V on each side would be plenty (even overkill maybe), since that would allow a 10V differential between grids at the max wiper positions?
 
For the project at the moment (meaning with the use of 8417 output tubes), I allowed for a maximum differential of just 4 volts between the output tube control grids, from one extreme to the other of rotation for the DC Balance control. Of course, that varies slightly with the current through the bias/balance circuit, which is based on the setting of the Bias control, but that is typically the range with the control biased for 40 ma of quiescent current.

The 8417 is such a very high Gm tube that any greater range within the balance control would make it very touchy to use and set. But this amplifier is still a development piece at this point, so I gave myself a fairly wide range to easily accommodate any of the tubes supplied. For lower Gm tubes -- and good basic design practice in general -- you can figure a total range equal to about 15% of the nominal bias voltage you expect to see at the grids will provide adequate range, while not being overly touching in use.

Dave
 
DIY Test Gear Thread?!

dcgillespie said:
Hi Jaz -- Thanks for the kind words. My lab equipment has evolved over the years with my original scope being one I built as a boy (Eico), and my first distortion measuring equipment being the really excellent Heath IM-5258 and IM-5248 THD and IMD analyzers respectively. My everyday workhorse signal generator is also a Heath IG-5218.

As I expanded into more critical work, and work requiring greater bandwidth than that of the basic audio spectrum (FM and MPX work), I upgraded my bench many years ago, so that now, my go to distortion analyzer is the superb HP 339A, and my DVM is also HP -- the 3466A -- whose calibration accuracy is traceable to the NIST. My scope is a workhorse Tek 475A. I also very much enjoy my HP 400 and 3400A AC and True RMS analog AC voltmeters as well. Finally, the remainder of my primary commercial gear is a Heath IM-4180 FM Deviation Meter, and a Fisher 300 Multiplex generator. There are other pieces as well -- an RF generator, Bridge, ect.

My custom equipment is my tube testers -- a power output tester for power tubes, and a Gm/AC Gain tester for small signal tubes. Good commercial tube testers are fine for culling out the shorted, dead, and gassy tubes, but without operating the tubes at real world voltage levels, give you only a very general picture of a tube's condition.

My power output tester is designed to operate a power tube under maximum rated cathode current under dynamic conditions, and the Gm tester tests small signal tubes under conditions outlined by RCA, Sylvania, and others in their tube manuals. With these testers, I can get a very precise picture of tube health as well as operating characteristic, that allows for extremely close grading and matching of power tubes, and small signal tube sections.

Thanks for the interest!

Dave
Click to expand...

Wow, that's a lot of gear :thmbsp: I know you are very busy, but perhaps when you get a chance, you could start a thread on the DIY test gear you have built, so some test-jockey wannabe like me can learn something - all I got now are some commercial test gear like HP & Tektronix, while they are good for everyday stuff, they aren't really designed to deal with tubes - certainly not at full operating voltage for prolong period of time :no:
 
dcgillespie said:
For lower Gm tubes -- and good basic design practice in general -- you can figure a total range equal to about 15% of the nominal bias voltage you expect to see at the grids will provide adequate range, while not being overly touching in use.
Click to expand...

Good deal, thanks. By the way, for me, this bias/balance approach presented a few challenges. There are several parameters to meet simultaneously--a) the right amount of balance voltage swing, b) current draw to run the circuit itself, and c) total grid leak resistance as seen from each of the control grids of the output tubes, looking back into the bias/balance circuit. :scratch2:
 
AC Balance - MO Mysteries, MO Questions, MO Answers

Just over two years ago, I commented on the design of the phase inverter used in these amplifiers, in yet another thread regarding one of these beasts. In it, I queried the value of plate and cathode load resistors used (33K and 22K respectively) and how different they were, versus the needs of the OPTs for maximum push-pull effectiveness. You can find that post here:

http://www.audiokarma.org/forums/showthread.php?t=396833&page=2&highlight=Bogen+MO-200A

The comment got little traction, most likely due to another AKer, who was nice enough to post an identical (in every way) schematic in the post following mine, except for revised values (now both 22K) used for these components in the inverter circuit. As a result, the concern was likely written off as a misprint, but I still had my concerns as:

1. The "misprint" was copied in both channels, and

2. The revised schematic provided sure looked like a Bogen schematic, but was copied from page 307 of a section of Super-Power Amplifiers in an (apparently) more generic book on this type of equipment. I have come to realize that schematics from such sources can often contain errors.

To make matters worse, in the official Bogen Owners Manual for these amplifiers (MO 100A and MO 200A), they list almost all of the parts used in these amplifiers, but conveniently left out the resistors used in and around the phase inverter circuits in the parts list, so there is no way to verify the values they labeled on their schematic.

And then of course, there is the actual unit at hand on my bench, which has original factory values of 22K installed for both components. It sure would seem then like there was in fact a misprint.

So what of all this nonsense? Well, the so what of it all came in my methodical approach in trying to get a handle on the distortion in this unit, where I first installed a true DC balance circuit in the output stage, then made the bias operating point adjustable, and finally then determined a good quiescent current to operate the output stage at. These efforts cut the initial distortion measured by a third, which now led me to install an AC balance control to investigate further distortion reduction via ensuring good dynamic balance. And that's where the nonsense all started up again.

In fact, even though Bogen installed 22K resistors for both the plate and cathode load resistors in the particular unit I have, minimum distortion was achieved with -- you guessed it -- a 33K plate load resistor, and a 22K cathode load resistor -- just as is shown on the official Bogen schematic -- and it dropped distortion fairly significantly, too. Now we're well under 1% at 100 watt RMS. But I gotta tell you, for minimum distortion to be produced with that large of a difference in these components, something had to be way wrong somewhere. Since the match of the output stage components and output tubes was now extremely good, it frankly only left the balance of the primary winding, and the feedback signal from that separate feedback winding in the output transformer.

Tests were run on the OPT and showed that the balance of the primary winding was within .1%, so there were no issues there, which I didn't think there would be. So the signal from the NFB winding was up for inspection next. Like all non-signal issues that NFB tries to correct, it also tries to correct for an unbalanced drive signal during Class B operation of a Class AB amplifier -- unless of course, the NFB signal supplied from the winding itself is not balanced due to poor coupling within the OPT.

The NFB circuit was disabled, where the ideal plate and cathode load resistor values in the phase inverter circuit immediately returned to a much more believable -- and normal -- 22.0K and 21.1K respectively to produce minimum distortion. Uh oh.

Remember, way back I mention that the NFB winding was not included in these transformers for any pretense of hifi performance, but simply to let the loaded output windings operate ungrounded as desired. Transformers with these windings when used in these kinds of applications often have less than stellar capabilities. From the results of the current testing, it appears that here as well, the use of that winding for NFB will likely hinder this amplifier from achieving its best performance.

There are alternate approaches to investigate of course, but my studies show that all modification efforts before me continued to rely on the NFB winding for the source of that signal, so it will be interesting to see if any alternate approach can return a worthwhile improvement in distortion reduction. After all, if the NFB signal is going to expend all its efforts to simply balance the push-pull signal (which of itself does reduce distortion), it can't be very effective in total distortion reduction in the manner which it could be.

This unit operates with a measured 17 db of NFB, which should reduce distortion by a factor of 7. With the NFB loop open and the improvement measures taken so far, distortion at 100 watts was clocked at 3.6%. Applying a 7 times reduction factor should bring distortion down to the .5% range -- but we're not there yet, still up at about .85%.

So, with (hopefully) an improved NFB design, and also not forgetting possible improvements from addressing the poor screen voltage regulation, we just might get this thing into hifi territory yet!

In the mean time, if any of you Aker's who have these amplifiers (100s or 200s) could verify the values actually installed for R7 and R26 (R8 in a 100) in your units (the resistor connected to pin #1 of each 7247 socket), it would be greatly appreciated!

Stay tuned.

Dave
 
Absolutely. Transformers designed for this service can be funny ducks. The most commonly used output winding in these units is typically the 70 volt winding. It may be that when that winding is loaded, the NFB winding works sufficiently well. But using more traditional impedance loads on a tap provided for that service, it is wanting. It is not uncommon that an 8 ohm tap (or there abouts) is not much more than an after thought design wise. Therefore, compromised performance on that tap may have just been considered acceptable. In some rather extensive work I did on these types of units a number of decades ago, I found that very much to be the case. The next step is to try the loaded winding itself.

Dave
 
With regards to regulation, have you noticed any issues with the grid bias voltage being unstable under full power conditions? I ask just because of the odd way its developed, through that capacitor tapped off the HV winding like it is. I've wondered if maybe it wouldn't be a good spot for a zener diode to nail down the voltage somewhat more firmly than just the resistive voltage divider rig it currently uses can provide.

I'll check the phase inverter resistors on my two 100a's later this evening and let you know what they have under the hood.
 
Thanks so much Gadget. The bias supply is definitely on my radar. I'll get into it when I get into the screen supply. For the moment however, I'm trying to nail down distortion based on circuit balance. Once that's done, then I'll go after it from a power supply standpoint, which will include all supply points to the output stage.

It will be interesting to know what you find in your 100As. As far as I can tell, Bogen didn't use serial numbers on these units, so you have little idea when one was built versus that of another, other than individual component manufacturing dates. This one also has a sticker on one OPT indicating a date of May 13, 1968.

Dave
 
Both of my 100a's have a 22k on pin 1, and an 820 ohm plus 22k on pin 3. Output from the cathode is at the junction of the 820 ohm and the 22k. Schematic that I have shows 22k/33k.

Something that may or may not be interesting. These were built about a year apart. One has a date of week 47 1965 on the power transformer, the other has week 45, 1964. The earlier one uses 150 ohm resistors on the output tube plates. The later one uses 39 ohm as specified on the schematic. The 39's look to have been replaced, and one is actually chopped out entirely. The 150's are smoked, but this was the amp that I got with a pair of 6L6G tubes installed in one set of sockets and 8417 in the other set. The whole bias balance circuit looks like someone took a blowtorch to it.
 
Thanks rpampt and gadget. I suspected that 22Ks would be installed. They are the theoretically correct values even though they cause notably increased distortion.

I'm testing out an idea regarding the phase inverter issue, and should have results to post in the near future.

Dave
 
Phase Inverter Mystery Solved

Sometimes it's hard to know how things came to be in a given design, and in this one, we may never know the real answer. It just seems strange however that in all of the official Bogen 100 and 200 schematics offered to date, the plate resistor of the phase inverter stage is given as 33K, while the installed value is 22K. This unit was produced for a number of years, so you'd think a correction or clarification would have been make at some point during that time. Add to it the fact that the labeled value produces far less distortion than the installed value, and the mystery only deepens.

While investigating this issue, the first test was to remove the NFB signal and see if doing so would cause the low distortion point to occur when the plate and cathode load resistors were balanced -- which it did. This proved that the issue lie with the feedback signal.

Tests were then run to check the effective coupling of the FB winding within the OPT to see if it in any way was contributing to the issue. I have seen this very sort of thing happen -- particularly at high frequencies -- where a poorly coupled feedback winding caused all manner of contorted phase distortion in the NFB signal at elevated power levels, which then disrupted an otherwise optimum AC balance under high level dynamic conditions. In this case, that was thankfully not the issue, but along the way, those tests returned other positive results with regards to frequency response, which will be addressed later in the project.

In this case, the problem was traced to a specific point of poor circuit design -- one which never should have been released by Bogen, but one which is also easily corrected.

The facts stack up this way:

1. The AF amplifier input stage of the 7247 tube operates with a static (cathode) bias voltage of .5 vdc. This is very low for the design, but was done to achieve maximum gain from the stage so that input sensitivity will still be acceptable after the NFB loop is closed. This loop reduces the sensitivity of the design by a factor of 7X, so all possible gain is needed to achieve a usable sensitivity in the final design of the unit when feedback was applied.

2. The design as offered requires 1.95 vac RMS to achieve 100 watts RMS power output. It is important to note that 1.95 volts RMS is equal to 2.76 peak volts.

3. The NFB signal is being taken from a dedicated NFB winding, which of itself produces 70 volts RMS when 100 watts RMS is being developed across the load intended output windings. This 70 volts is then divided down by the NFB network that effectively consists of a 56K NFB resistor, and 1200 ohm cathode resistor of the AF amplifier stage. This network then is applying 1.47 volts RMS of feedback signal into the AF amplifier input stage. Again, it is important to note that this 1.47 volts RMS of feedback signal is equal to 2.08 peak volts.

4. When the input signal swings to its most positive point then, a positive 2.76 peak volts is applied to the grid of the AF amplifier stage (from #2 above). Against this, is the positive .5 vdc of static bias (from #1 above), and the positive 2.08 peak volts of feedback signal (from #3 above), that add together to produce a cathode voltage of positive 2.56 peaks volts at the time that 2.76 positive peak volts is applied to the grid. Uhhhhhhhhh, that means that even before full power output is achieved, the input grid has been driven positive, causing a highly distorted signal at the plate of the AF Amplifier stage. Bottom line? You can't do that!

5. This signal is highly asymmetrical, and as applied to the inverter stage, requires it to employ grossly unbalanced plate and cathode resistors DURING THIS TIME to effectively apply a balanced signal again to the output stage.

Again, this should have been caught well before release. But, stranger things have happened.

The answer is to apply more static bias to the AF amplifier stage, while altering the cathode circuit to still allow the stage to maintain maximum gain.

A quick test of this returned a Bogen that required perfectly matched 22.0K plate and cathode resistors in the inverter stage to return the lowest distortion under ALL levels of power output -- AND -- when combined with the optimum quiescent bias level in the output stage, returned a 100+ watt RMS THD level of just .20%. There is more work needed to reliably achieve this distortion level, but suffice to say, this represents nearly a 90% reduction in distortion from the original, properly operating design.

So, progress is being made. It will slow now with the Holidays upon us. But this should certainly wet the whistle of you Bogen owners out there. Frequency response should be able to be greatly extended beyond that of the stock design as well.

Stay tuned!

Dave
 
Hiya Dave,

Well I knew you would find something and my impression of you while high already is much higher now.

I looked at my 100A's and I have the 22K's on Pin 1 and 22K's also on Pin 3 .. Both appear to be stock in every case.

Frannie
 
Hi Frannie -- Thanks for the kind words. These units will still be somewhat bandwidth limited of course due to the tests that covered that element of this project early on. But, the bandwidth has the possibility of being extended somewhat now, and within that bandwidth, performance should be markedly improved. I note that others who have heard the stock design have called it lack-luster -- and with all that has been uncovered, it is easy to see why. Of course when I'm finished, I'll supply a final schematic for all you Bogen heads to go after it!

Dave
 
You must log in or register to reply here.
Back
Top Bottom