dcgillespie wrote:Lots going on now!
Greg -- One of your previous posts indicates you were getting over 90 watts RMS out of the unit before you modified yours with the power scaling circuit. You also provided some power supply readings at various power levels. This answers many questions!
1. It shows that the original OPT as a matter of size can in fact pass the power level the tubes and power supply are capable of -- even if only at mid-band frequencies.
2. It also shows that the proper load (OPT impedance) is there to sink the current that the tubes and power supply are capable of sinking/supplying themselves.
3. It shows that the power transformer is rather weak, as a 60 volt drop on the main output from a 480 vdc quiescent condition indicates a regulation level of 12.5%.
This is all much closer to what the "evidence"indicates should be produced, although frankly, your power output reading seems somewhat high. With 420 vdc on the plates, and only 318 vdc on the screens, the tubes simply cannot pull enough current across a 2500 ohm load to develop the kind of power you are indicating across an external load. A 90+ watt power level WOULD be developed at the plates of the output tubes in this scenario, but after typical OPT losses, the power actually delivered into an external load would be around 75 watts RMS maximum. A number of things can come into play to create power reading errors, being the accuracy of the load used at operating temperature, connecting lead losses versus where the measurement is taken, and meter accuracy, both with respect to calibration and frequency response.
As for the delayed screens, the OP may have used a tube due to my warning to never have screen power applied to warm output tubes without the plate power applied also. To do so would ruin the tubes rather immediately.
Regarding the article, I'm glad you found it! I believe I also have another article comparing a traditional full wave design with SS diodes, versus a full wave SS voltage doubler. I will look to see if I can find that one for you as well. Again however, due to the considerable winding resistance difference in the power transformers of these two designs, the SS doubler does win out.
As an example, my primary system is of my own design, employing PP 6550s in each channel, with each channel capable of 75 watts RMS from 20 Hz to 20 kHz at this power level. The worst case scenario is 75 watts RMS at 20 Hz, where each output stage is drawing ~ 375 ma, or .75 amp with both channels driven. These amplifiers are powered by a common power supply, with common active regulation, which supply the output stages a steady 485 vdc regardless of signal level in either channel.
This regulator is supplied from a raw supply consisting of a full wave voltage doubler configuration, with two 5A high speed diodes, two 400 mfd 350 v doubler caps, and a power transformer rated at 1 KW. With a 203 vac secondary, it is capable of delivering nearly 5A RMS current flow. Against this, the complete amplifer under worst case conditions requires .8 amp to produce full power in both channels. But to provide this current level from the power transformer, this requirement must be doubled due to the doubler's action, and, the current level must be increased even further because of the cap input nature of any doubler design. In this case then, the effective draw from the power transformer is just over 2.25 amps at full power under the conditions described above.
At quiescent conditions, this supply is delivering 535 vdc to the regulator. Under worst case full power conditions, power to the regulator drops to 517 vdc. With only an 18 volt drop, this means the regulator itself is supplied with very well regulated voltage, with regulation better than 3.5%. With this kind of performance, the regulator's ability is enhanced, as the drop across it is minimized, as is dissipation, and performance is ensured at line voltage variations of +/- 5%. This kind of performance could not be delivered with a traditional full-wave design, as the winding resistances for the power transformer of such a supply -- while low -- would still be significant compared to that of the transformer I am using, where total winding resistance in ohms can be counted on one hand.
While my example may be extreme (requiring soft start circuits, etc.), it still illustrates the capability that the full wave voltage doubler is capable of. When that is considered along with the ease of transformer design it creates, it is no wonder that virtually all of the high power amplifiers converted over to this type of design when dependable SS rectifiers became available for this type of circuit.
Dave
Dave,
Forgive me for a typo that I went back and corrected. The 90+ watts were peak and not RMS. Those were taken with the amp full up, lots of distortion, etc. The RMS power on that revision was 51 watts RMS. That revision had a cathodyne inverter (12AV7) and a 6BR8 in the preamp, whereas the current revision is using a LTP (12BZ7) and a parallel 12AX7 and a 5879 in the preamp and gets 58 watts RMS before the power scaling was added. It sounds better too with the current revision. I haven't checked the power output since adding power scaling because I know I have a problem with the power scaling circuit. It functions fine and reduces power as it is supposed to, with the bias tracking the power reductions, etc., but the B+ MOSFET is dropping around 40v and should not drop so much. Kevin O' Connor (power scaling circuit inventor) recently emailed me regarding some info I supplied him about it and said I have a couple things hooked up incorrectly, so I'll be delving into that after finals (next week) and an out of town trip, hopefully, although the next term of school will be starting up too. I will have to find some free time because I want to finish the Bogen project and get back to my Vox AC100 project which is 3/4 of the way done.
I would suspect that part of the previous sag was due to a problem I had with the first dropping network in the amp that I have since fixed. Part of it is likely due to a lower quality power transformer too I would guess. It isn't all that large for a 50+ watt amp. Once I get this amp completely functioning correctly I will take all the readings again and post the results and then we should be able to tell about the transformer spec. I plan to add a power scale disconnect switch to the amp so I can get readings without it in the circuit too.
Your example amp sounds interesting! I've only done guitar and bass amp projects for myself and fixed a lot of amps for others at this point, but I do have some hi-fi amps to work on. I have a Dynaco Stereo 70 and a Scott 299B to overhaul among others. I also have a 1964 RCA console stereo to overhaul, but it works perfectly (except for FM) still on original tubes and caps so I haven't bothered and use it often. I would be very interested in seeing the other article that you mentioned regarding the SS voltage doubler vs the traditional full wave SS setup.
Greg