Magnavox Flea Power: Getting More Out Of The 8600 Series - A Lot More!

[derekva]

Yes, thanks to Dave for another great tutorial and taking the time to set me straight on a few things. But I am sure there are many that are not going to shell out $30 for the replacement OPTs, in which case, the second best thing to do is follow Dave's instruction to re-bias the 6BQ5's and live with the reduced performance.

On the other hand, for less than $150 in parts (including tubes), you can build a very respectable SEP amplifier that has some solid engineering behind it from scratch. If you wanted to spend a tad more, you could be a little more boutique and easily compete head-to-head with a Decware Zen.

-D

 

[opa1]

What is the wattage of the cathode resistors and volts of the PT?

 

[dcgillespie]

Primo -- I have seen some of the various offerings that include a LM317T configured for constant current operation. This is one of those modifications that would surely make good sense, as it acts to force Class A operation.

If you are trying to run the tube at a lower than optimum quiescent current for tube life (what other reason would you? No one would really ever run a tube at higher than optimum.....), you have inherently moved the tube away from Class A operation at the lower end of the load line. In a push-pull amplifier where there is likely plenty of reserve power available, concern for the operation at the extremes of the load line may not be so important. But in a flea installation where every portion of every watt counts, the performance along the entire load line becomes crucially important.

So what does a regulator do in an over-biased scenario? If the tube gets driven into early cutoff because of using a simple -- higher than optimum -- cathode resistor, then replaced with a constant current regulator set for the same standing current, the regulator would automatically try to compensate for the lower current draw produced by the non-linear operation by decreasing the bias to restore the original current draw level. It sounds good -- but how effective is this?

Remember that such a regulator has to be heavily bypassed with a large cap -- otherwise there would be zero output at the plate: For every cycle where plate current increases from a positive grid signal, the regulator would simply act to increase bias to restore the original current level. On negative grid signals, the opposite would happen, meaning that there would be zero output from the applied grid signal. Therefore, to prevent this from occurring, the regulator must be heavily bypassed so that the regulator can only respond to AVERAGE increases or decreases in current flow. But then -- the cathode resistor is heavily bypassed as well. So in this scenario, the regulator -- as an active device -- is providing some technical improvement, but since the cathode bypass cap is largely the control element in the cathode with regards to dynamic operation, the real improvement achieved with the regulator is likely minimal.

Against that, biasing the stage properly at the optimum quiescent operating point (by any means) produces an inherently natural action of Class A operation, with very little if any average current change naturally occurring through the cathode biasing element. Power is maximized, and distortion minimized. Fixed or constant current biasing arrangements would have little if any benefit over conventional resistive cathode bias in this scenario.

Now if the regulator is used to bias the output stage to the proper quiescent current, then it would aid in achieving that setting for (theoretically) any tolerance range of tubes better than a simple resistor can. But at this point, that's about the only benefit I can see.

Still, I like to consider myself as open minded, so I will play around with the concept in the near future. Thanks for adding the thought!

Dave

 

[dcgillespie]

Cathode resistors are all .5 watt. It is the original Magnavox power transformer -- if any knows the voltage of the HV winding, please chime in. Otherwise, I will measure mine later on today.

Dave

 

[jazbo8]

if any knows the voltage of the HV winding, please chime in. Otherwise, I will measure mine later on today.

Looking at several 86-xx schematics, the PT appears to be 230-0-230, which matches up well with the voltage that you indicated on your schematic.

 

[Dandy]

Dave, I just want to add my thanks for your tuition in this thread, and the schematic. Much appreciated!

 

[Wharfcreek]

Dave, how much 'wiggle room' is there in the secondary of the PT? The '500VAC' output being discussed..... is locked in stone? I'm looking at another chassis I have that is a SE EL84 amp of some kind....... and the secondary on that PT is about 560 (unloaded). I'm just wondering if that make it 'unusable' in terms of this project, or if that's 'close enough' that perhaps I can 'tweak' some in the filter stage to make it work. Thanks, Tom

 

[Tom Bavis]

With a higher B+, the idle current might have to be a trifle lower to stay inside the dissipation rating. So the optimum transformer impedance might be more like 6K. Once you decide on the DC operating point, you can look at the load lines for 5K, 6K, etc.

 

[dcgillespie]

The neat thing about a properly designed singled ended amplifier, is that the current drawn by the amplifier is virtually or nearly constant from the power supply, regardless of power output level. Therefore, if the B+ is a little high, and you want to stay with the basic amplifier design as is, you can in fact also just use additional resistance or filter sections to drop it down to the required voltage. SE designs need very well filtered B+ voltage to eliminate PS noise from the output, so having a slightly higher B+ voltage can be of some benefit if it is put to good use in improving PS filtering.

Dave

 

[kward]

Dave, would a fixed biased design offer any more power, or offer any other benefits over a cathode biased design?

Also, how did you select the amount of NFB? Was that just what was there in the original design? (Edit: Yes I see now you kept the feedback level the same as the original design when moving to the new transformers with the 8 ohm secondaries).

 

[for_p1]

Dave, would a fixed biased design offer any more power, or offer any other benefits over a cathode biased design?

Also, how did you select the amount of NFB? Was that just what was there in the original design?

Fixed bias only allows higher power in push-pull circuit by bringing tubes deep into class B operation.

For SE circuit it allows exclude high value electrolytic capacitor from signal path, and thus reduce possible distortions.

 

[dcgillespie]

Agreed. The only increase in power output that would be achieved by going to fixed bias would be due to regaining the DC voltage drop currently lost across the cathode bias resistor. But because the 6BQ5 is such a high Gm tube, the voltage to be regained is so small that the additional power output achieved would be insignificant.

Dave

 

[Wharfcreek]

Dave, you mentioned that you have been using this with your Cornwalls as a means to run it through it's paces. I'm lucky enough to have a set of those myself..... and would no-doubt want to hear this little amp 'at it's best' (so to speak) when driving those highly efficient speakers. However, I have to say that when I first got on a bit of a 'kick' with 'flea' amps, my thoughts were centered around some relatively small speakers as well. In fact, I had attended the AK Fest in Detroit where I saw a set of what I believe were some Madisound 'Transmission Line' SE speakers that had a mere 5" full-range driver in them. At any rate, I began to envision a system built around some speakers like these along with an amp like the little Maggie, a decent pre-amp though again VERY simple, and built for some focused vinyl listening. I still have that vision, but I think the changes you've made in the amp will result in far greater detail. I'd thought about the impedance mis-match...which I figured I'd solve by using 4 ohm drivers in the full-range system. There are really a LOT of these available now..... including some 8", 10", and even 12" versions. But while the outcome here seems a bit focused on 'power'..... the unsung hero here may be the clearly obvious results in reduced distortion....which seems to me to be every bit as significant as the power gain. TSD

 

[jazbo8]

<snip>constant current biasing arrangements would have little if any benefit over conventional resistive cathode bias in this scenario.

While RH Amplifier and others have been touting the benefit of using CCS in the cathode in recent years, Dave as usual has done a masterful job in debunking the benefit of such "feature", just stick with the good ol' resistor (properly bypassed of course) and be done with it.

 

[Leland47]

Wonderful thread. Boy, do I have a lot to study and learn..!

Thanks so much,

 

[dcgillespie]

mis-match...which I figured I'd solve by using 4 ohm drivers in the full-range system. There are really a LOT of these available now..... including some 8", 10", and even 12" versions.

This is surely one of the things I learned throughout this project. I guess being hopelessly stuck in the idea of "vintage", I think of such systems in their entirety that way -- including the speakers -- with the exception of modern sources of course. Vintage speakers throughout the heyday of stereo were largely rated at 8 ohms, with 16 ohms being common throughout the 50s in the earliest days of the audio revolution. No doubt that some modern speakers today have moved to 4 ohm drivers to make the most of SS designs. It would be nice to think however that manufacturers did so purely to serve the 4 ohm Maggie crowd.

If the design is to accommodate 4 ohm speakers as well, then the logical move would be to use a 5 watt (or greater) Hammond Universal SE OPT to (hopefully) accommodate all possible speaker impedance levels. I did not try it, but I have a sneaking suspicion that the original 4 ohm transformers will likely saturate at the elevated quiescent current levels (over stock) required for proper output stage operation -- except that those larger Hammond transformers won't fit on the 8600 chassis -- not neatly anyway. If you go that route, you would almost surely need to do so as a scratch build, as such an effort with the 8600 chassis would end up turning it into a contorted, over stuffed contraption, coming off as too cute by half. As I have always stated and has been echoed by others, at what point then does the Maggie cease being a Maggie? Heavens knows that question has been asked by about a half a bazillion Dynaco owners as well..........

The ultimate answer would be to go to a scratch build, and a truly high quality OPT, with designated 4-8-16 ohm output taps. Of course, any of these options would require reworking the NFB loop to maintain the original FB levels and proper stability. One thing is certain however, and that is that amplifiers of this type obviously have a huge following -- so large as to accommodate anything from those using the basic 8600 as originally executed, to those with all out cost no object scratch builds. In that vein then, my project here is but a pebble in the ocean! Hopefully however, it has shed some light on the all important areas of maintaining stability, optimizing loading/impedance match, and using proper output stage quiescent current levels in achieving the best performance outcome.

Dave

 

[Wharfcreek]

Dave, in a PM with a fellow AKr, I dubbed this the DG-SE1. I can't say 'when' it no longer became a 'Maggie'...... but where that train left the station as an old Maggie Steam Locomotive, I think it arrived as the DG-SE1 Diesel Express!! FWIW, with much gratitude! Tom

 

[wa2ise]

Dave, would a fixed biased design offer any more power, or offer any other benefits over a cathode biased design?

.

Even if you did get a little more power, you'd never notice it. The ear can't tell increases less than a dB or 2. I've actually decreased power output in other types of amps a little to avoid stress on tubes, knowing I'd wouldn't miss it, and get better reliability.

I can't say 'when' it no longer became a 'Maggie'...

I'd say that it is no longer a maggie when you change the output transformers.

 

[Wharfcreek]

Well, I'm on my 4th or 5th complete re-read of this entire thread. While I have to admit that the discussion on 'load-line' got a bit over my head, I'm still intrigued with the idea of the LM317 as part of the bias circuit.......AND.... I note that S-Peterson had questioned about the use of SS rectifier vs the 6CA4. Dave, if you decide to fiddle with that voltage regulator, maybe you could also include some consideration of the SS rectifier as well. As Scott indicated, less current draw from the PT, potentially a few more volts to work with.....? I know it's not the tube-purist amp any more....but perhaps could be the DG-SE2? lol. TSD

 

[jazbo8]

I'm still intrigued with the idea of the LM317 as part of the bias circuit

As Dave already explained, there is un-likely to be any benefit to use the LM317 CCS in a SE output stage, so what's the motivation? Fashion?