In the process of rebuilding my old 300B amplifier, I decided to make a new filament power supply. It all came up around the components I had at hand, so it could’ve been improved but that meant extra cost:
I have a pair of custom-made JMS transformers with multi-taps secondaries. This helps me tweaking the right output voltage. Anyhow, any 15V transformer would do. Perhaps you want 14V to ensure you don’t dissipate too much on the filament regulators (e.g. Rod Coleman regulators)
I used my flexible LT supply PCB which allowed me to build this in less than 1 hour. I also used some existing chokes made in the UK by “Spirit” which are ok for this purpose. The Lundahls are in use, so can’t reuse them:
I used some SOT-128-2 schottky rectifiers but any other should work as well. Resistors are wirewound and the CMR choke is what I had in stock as well. A simple 15mH/3A should do fine.
The output measured well at 9V with a 6R load which dragged 1.5A. A bit more than the 300B but should be a good indication of performance. Also ripple level is good at 2.5mVrms. The rest will be cleaned up by the regulator itself 🙂
Hi Ale,
Your raw DC supply seems to include a bleeder resistor, R5, that never seems to show up in Rod Coleman’s raw DC supplies. The inclusion of a bleeder has been something I’ve thought about, as I always include one in high voltage supplies, so could you please comment about that?
Also, is the LED simply an indicator that the unit is *on*? And, how did you pick the choke?
Is your board available for purchase, and if so does it include the SMD Shottkys soldered in? Price? Shipping to USA?
Thanks and Best Regards,
Robert Chambers
Corinth, Vermont USA
c
Hello Robert,
I believe that the reason that Rod Coleman’s Raw DC Supply does not include a bleeder resistor is that it is capacitor-input DC supply as opposed to a choke-input DC Supply. At startup, the Coleman Filament Regulator is slowly feeding current into the high impedance of the cold DHT filament. With a capacitor-input supply at startup, the raw DC voltage supplied to the regulator is only slightly higher during this warmup phase. However, with a choke-input DC supply at startup, there is not enough current flowing through the choke so the raw DC voltage will soar to 1.4142 * Vrms of the transformer secondary instead of the roughly 0.9 * Vrms of the transformer secondary during normal operation after the DHT filament has warmed up. Adding a properly sized bleeder resistor to the choke-input DC supply keeps enough current flowing through the choke to keep the raw DC voltage at roughly 0.9 * Vrms of the transformer secondary.
Best regards,
Paul
Hi Robert and Paul
As Paul explained, it’s a choke-input supply hence you want the minimum load current to prevent the output voltage from raising. Nevertheless is a good practice to overrate cap maximum voltage in case of any issues. I also do the test of removing the load and monitoring the potential “ringing” of the supply with an oscilloscope in DC-mode. I previously checked with PSUD2, however is always good to double check that the supply has a nice response to step changes in the load and no signs of ringing.
There is a nice chapter in Morgan Jones’ Valve Amplifiers Fourth Edition which covers well this topic. In particular, this supply is not a pure choke-input, it’s rather an “intermediate-mode” PS. It uses the voltage “tunning” capacitor C2 in my board.
Hi Robert
I have a few made for myself only. I’m not convinced in offering these as I may end up with tons of questions regarding the supply design, which I don’t have the time to support am afraid. Also don’t have the time to pull together a BOM. If you’re still interested in some for your own work (and you know well what you’re doing with the PS), then drop me an email.
Thanks
Ale
Hello Paul and Ale,
I thank you both for your replies…it’s helping me to think about how to make this more of a cookbook approach to sizing a bleeder resistor for a choke input supply.
But being an (old) ME, not EE, I am challenged and slow as you may gauge by the time it took for me to reply to your posts. 🙂 I keep thinking about this and re-reading your replies, I still cannot come up with the cookbook recipe, although I think Paul has explained it well…I will keep on trying.
Very Best and Thanks,
Robert.
Hey Ale,
How are you? It’s been a long time we haven;t crossed.
I have just build a replice of the Kumuro 300B amp just like the original found in the internet with 2 variations: the 6C4C replaces the 300B and its filament is DC heated. At the moment the DC heating is a simple R-C-L-C filter and I get about 2 mV ripple peak-to-peak. The caps are EPCOS 22000 uF (17mR ESR). The choke is 22mH, R= 3.3R (it eats quite voltage but limits the current). Why do you get such high ripple with one extra LC filter? I have not measured the inductance of my choke (as one needs a DC superposition and it gets too complicated with basic equipment), however I am confident the value I have estimated is not overstimated (having used a pretty low relative permeability). By the way the 6C4C works at 345V/55mA and it delivers 7W rms (yes!) into 5K. The sinusoid at that power level looks perfect… 🙂 …I would guess 2-3% THD at worst. The sound is nothing less amazing, especially when paired with the 92 dB/1W TQWT I have built with the 8″ SB-Acoustics fullrange (you can find the project online). I know that 19W dissipation but the 6C4C doesn’t seem to bother. It has been working for several hundred hours now. Not an issue and it’s quiet, despite the high efficent speakers.
Cheers,
45
Ciao Paolo,
So great to hear from you! How’ve you been?
Interesting experience you share. The Komuro design is very standard and has 2 driver stages which is not my cup of tea. What driver are you using? Worth replacing 2 driver stages with a single one in my experience. Can’t comment on the 6C4C running that hard, not surprised as have used many Russian valves pushed beyond their limits. If they still are after several hundred hours, then great.
The PSU filtering can be tricky, worth modelling on PSUD2 and see if you have some poles not dampened properly. Generally you can model this pretty accurately despite not having a precision measurement of the choke inductances. You can play around the step response on the simulation and see how much additional tiny series resistor may be required or tweaking your caps. There are a couple of very good tutorials out there on how to do this job with PSUD2.
Cheers
Ale
Hey,
I have been okay. Thanks.
I gave it a further a thought and I suspect that the common mode choke you use is actually filtering the differential signal thorugh its leakage inductance only which should be in the microH region.
The 6C4C at 19W (I can swear) is not running hard at all. The bulb temp is also at very acceptable level. The amp has been running every day for over 3 months now. The valve tests the same on the Hickock.
For myself, I am going to build it as a 45 PSE and use the Coleman regs. 320-330V at 30 mA for each tube with 3.5K. This should give 5.5-6W without enetering positive grid. I definitely prefer the higher voltage + higher primary load solution for most of my amps. Distortion is lower and effeciency goes up. I still have the 3.5K NP Acoustics unused. It’s a pity the business is gone. Their website is down. I will start it during Xmas holidays, I hope. The amp with the 6C4C insted will be probably work with the 300B and use the 6C4C as the cheaper option. Enresto is designing a board for the front-end with its PSU and boards for the DC filament supply. The idea is that only the the power stage and its anode supply will be wired point-to-point.
Oh yes, the 45 PSE is a wise choice. Hopefully you have enough NOS stock of them or alternative invest your saving in the EML ones, which are fantastic BTW.
Sounds like a great project for Christmas!
Cheers
Ale
I missed you question abot the front-end. It’s the same as the original except the 6J5 has been replaced with the 6SN7. I have designed a better anode choke which is 50H minimum with small signal applied. I think we will go for the PCL82 for the kit. Ernesto has a good stock of quality manufacture. I have seen that the commercially avaialble Komuro uses the 6BM8 too. The one I will build for me, I likely will use the 12AY7 and the 6F6GT. I have a nice batch of 1951 Fivre’s with brown base and grey coated glass. I have thought about the 2-stage solution but there are not many tubes for it and will need some compromise. High enough gain + 20mA + anode current is not easy. From what I can listen the 30-stage solution is not a dowgrade with full DC coupling. I am also using a DC-link cap for the cathode bypass of the power tube with cut-off at around 10Hz in order to have faster recovery on large transients (basically 5xRC = 80 ms). The anode current of the 6V6 in the prototype is actually lower than the original at 22 mA. This way I could leave the orginal 1K cathode resistor. The full bias is made of 10V across the anode choke, 55V across the cathode resistor which is on the negative side of the filament and 3V volts to the centre filament for a total of about 68V. Being a stereo amp the 5U4G works for both channels. IT’s easy to adjust the anode supply being a pseudo choke input supply.
Cheers,
45
Ciao Paolo
Yes, there are so many ways to skin a cat as they say. Personally, I prefer a high-gm pentode triode strapped as a driver for 1 stage only. I used the D3a, 12HL7 and 6J52p with a lot of success. Even with a DHT, I’m using te 47 with an input SUT and DC coupling into the 300B with fantastic sound, I’m not moving out of this topology for now.
Cheers
I know, not trying to convince you. I don’t like the sound of frame grid tubes used as power tube drivers and even worse the SUT solution. It’s a matter of preference.
Also running power tubes at higher current and lower anode load doesn’t bring much benefit in terms of real distortion. So far 420v/70mA with 5K has worked better with all the 300Bs I have tried.
45
Here at home now, from my notes, the last measurements on the 300B-Z resulted in 1% at just under 7W and 3% at nearly 11W. The output transformers were the same we are using now. Geometry is the same of that I designed for your 814 amp. It’s only made on a smaller 36×50 core and has a bit better primary-to-secondary insulation. The FR with both 300B and 6C4C is dead flat from 10Hz to 42KHz. Nice square wave even at 20 KHz. The output stage with 300B running at 420V/70mA with 5K, I think you should try and not rely on sims that rarely get distortion right. I’ ll send you some pictures when we make the final version into a nice chassis. 😉
Cheers,
45