Here is my first attempt at a mini (actually an micro-amp) using the Russian equivalent of the 6BM8. I have an octet of tubes on the way, and a a pair of bitty output transformers in fab at Edcor. The boards are in process at PCBway. I plan to power the amp using a DC-DC converter run by a 12V, 3.5A switching adapter. The amp takes its inspiration right from the Sylvania data sheet, with the power pentode being running full tilt at its 7W dissipation rating. I am running the input triode at 1ma plate current with an LND150 current source load. This will go into a proletarian bud box (so not all that pretty), with transformers on the bottom to keep the thing from tipping over. They will drive a set of Klipsch speakers that are actually pretty efficient (they have no problem making lots of noise with the 20W/channel Class A amp currently driving them). The amp will get a full characterization on the bench for frequency response and gain-phase, as I have a gain-phase analyzer available at work that goes up to 5 MHz. The values on the output Zobel network will be adjusted so as to keep the output gain from going back up through 0dB (gain margin as well as phase margin).
Attachments
Last edited:
It is always nice to try ones's first attempt at something different.
In your case . . . a mini.
Have Fun!
Let us know how it comes out.
In your case . . . a mini.
Have Fun!
Let us know how it comes out.
I'm curious to see how loud it will drive my Klipsch speakers... I did another teeny amp I christened the "Mighty Mite", also powered with an SMPS. It used a pair of 6LR8s with the big pentode trioded. The thread for that effort is here -
https://www.diyaudio.com/community/threads/mighty-mite-single-ended-amplifier.120804/
https://www.diyaudio.com/community/threads/mighty-mite-single-ended-amplifier.120804/
Last edited:
You have a Zoebel network across the OPT outputs (C7A, R14A). That is something I have wondered about for amps I have built, especially if they can sound a bit shrill at times. How did you derive the values?
Some of the parts in the schematic might need to be optimised for the output transformers, or is this a tried and tested combination?
Some of the parts in the schematic might need to be optimised for the output transformers, or is this a tried and tested combination?
OldHector - the Zobel values in place are merely placeholders until I can get the finished amp hooked up to a Gain-Phase analyzer. I will then adjust the Zobel network until the frequency response crosses below zero and stays there. Without the Zobel. the gain can cross zero, then swoop back up through zero again at an unspecified higher frequency - this is the very definition of conditional stability. With a properly adjusted Zobel network, I can make the frequency response cross below 0dB and stay down. This is a condition required for gain margin, which is as important as phase margin to get a stable amplifier.
This is not a tried and tested combination, but a work in progress - I may have to adjust the cathode bias on the output tube to be compatible with the 40 mA current rating of the Edcor 5W output transformers.
This is not a tried and tested combination, but a work in progress - I may have to adjust the cathode bias on the output tube to be compatible with the 40 mA current rating of the Edcor 5W output transformers.
Last edited:
I tried something similar with the 6DX8 (yea, different animal) and was a bit disappointed. Enough power to drive Heresy's to moderate listening levels, but not as linear as I had hoped since they were video amps. The 6BM8 is well regarded in older equipment from what I understand. I look forward to your assessment.
My first tube project was a small SE amp with 16A8, the European 300mA series heater version of 6BM8. This was over two decades ago but I can still remember the raw excitement when I hit the power switch on the breadboarded mono prototype, watched the heater light up and then heard music from the test speaker. I guess that was the equivalent to a junkies first shot of heroin, no thoriated tungsten transmitter tubes or thousand dollar transformers has ever managed to top that...
Edcor just shipped my transformers - I wasn't expecting them to cough up until the end of July - a pleasant surprise, for sure...
Maybe they haven't had so many customers as of late, as they jacked their prices up past the stratosphere... They used to be the cheap go-to, if you were willing to wait.
Maybe they haven't had so many customers as of late, as they jacked their prices up past the stratosphere... They used to be the cheap go-to, if you were willing to wait.
Hi Wrenchone.
I am also looking forward to the completion of this build and may well build something very similar if not identical.
I have also been intending to build a 12 volt tube amplifier. I was going to wire it for a 100Hz or 140Hz vibrator using a 4 or 5 pin socket and plugging in one of me electonic vibators. The idea is it could work with a vintage vibrator or one of my new ones. (redesigned this year). Here is a link to the old version https://sites.google.com/view/kens-schematics-2/power-supply-vintage-radio-2024
I know I'm going off topic here but I have found many small mains transformers work well at 100Hz and some even very well up to 400Hz. When the frequency increases the voltage into a winding can be increased as the core flux density goes down.
I was running 12V (DC supply) into a 7.5V winding (BTW 7.5VAC has a peak voltage of 10.6V) with most acceptable results.
I am also looking forward to the completion of this build and may well build something very similar if not identical.
I have also been intending to build a 12 volt tube amplifier. I was going to wire it for a 100Hz or 140Hz vibrator using a 4 or 5 pin socket and plugging in one of me electonic vibators. The idea is it could work with a vintage vibrator or one of my new ones. (redesigned this year). Here is a link to the old version https://sites.google.com/view/kens-schematics-2/power-supply-vintage-radio-2024
I know I'm going off topic here but I have found many small mains transformers work well at 100Hz and some even very well up to 400Hz. When the frequency increases the voltage into a winding can be increased as the core flux density goes down.
I was running 12V (DC supply) into a 7.5V winding (BTW 7.5VAC has a peak voltage of 10.6V) with most acceptable results.
I'm doing a 60kHz take on your vibrator setup with a modified push-pull DC-DC converter. I'll see your vibrator and raise you....
I did the usual 6EM7 mini amp with great results. I wonder how it compares to this amp? I went with a voltage doubler as I’ve never liked the idea of a switching supply in audio, especially in a tube situation. As far as OPTs, I’ve had good luck with Chinese units from HiFiEquis on AliExpress. They are a range of transformers available for very cheap but the shipping can kill you! It’s more than the transformers! I think they’re nice than the mid range Edcorel its crazy how good a little amp with a watt or two can sound if you don’t push it beyond what it is meant to do.
I have absolutely no problems with SMPS in a tube context. Properly shielded, they deliver the goods. As I've been designing SMPS for a living for over 40 years, I know whereof I speak.
I hope it works out well for you. Will you use a micro?
I have found the programmable OFF time is very important when striving for the best results. To some extant the ability to limit the slew rate with the gate resistor can also be beneficial. My new PCB uses only through hole components and uses a DIP 8 micro in a socket so those not into programming can simply plug in a new I.C. The DIP switch allows 8 programs per chip to be selected. It is a push-pull device. The PCB is 30mm x 60mm.
In the tube circuit, use a vibration source to convert 12V into a power supply voltage for the tube to work effectively.
Obviously you know how to make an SMPS that won’t make the amp it’s powering sound like an old 56k modem connecting. 40 years will tend to do that. Somebody building his first or second tube amp might not have as good a result. Buying one on e-bay doesn’t guarantee quietness, either. One tends to get what they pay for - if it was just intended to run LED lights it could be a nightmare.
Tubekwk-
No micro is getting near the SMPS. I'm using a current-fed push-pull topology and generating opposite phase square wave signals using a CD4047 astable multivibrator. These get some delay applied using an RCD network to generate about 250ns of overlap between phases, which keeps the current feed inductor from going crazy (with a current fed topology, you want overlap between phases rather than dead time). These then get applied to TPS2828DBVR driver chips . These chips have a Schmitt trigger input that squares up the RC delay signal and drives the push-pull mosfets. It's a pretty simple circuit, In principle. It has worked very well for me powering a 13DR7 SE headphone amp, a thyratron blinker circuit, and an 8-9W SE amplifier using a 6AH6 pentode front end and 6L6GC finals, with plate-plate compensated partial feedback between the input and output pentodes. All the amplifier circuits have been wrung out and stabilized using a gain-phase analyzer I have available at my workplace.
Using a push-pull topology instead of a flyback makes good use of the voltage available from medium -to-high powered 12-24V switching adapters. It moves the safety interface out to the switching adapter, and eliminates the weight of the power supply XFMR (though you still have the weight of the 2 output XFMRS to deal with).
No micro is getting near the SMPS. I'm using a current-fed push-pull topology and generating opposite phase square wave signals using a CD4047 astable multivibrator. These get some delay applied using an RCD network to generate about 250ns of overlap between phases, which keeps the current feed inductor from going crazy (with a current fed topology, you want overlap between phases rather than dead time). These then get applied to TPS2828DBVR driver chips . These chips have a Schmitt trigger input that squares up the RC delay signal and drives the push-pull mosfets. It's a pretty simple circuit, In principle. It has worked very well for me powering a 13DR7 SE headphone amp, a thyratron blinker circuit, and an 8-9W SE amplifier using a 6AH6 pentode front end and 6L6GC finals, with plate-plate compensated partial feedback between the input and output pentodes. All the amplifier circuits have been wrung out and stabilized using a gain-phase analyzer I have available at my workplace.
Using a push-pull topology instead of a flyback makes good use of the voltage available from medium -to-high powered 12-24V switching adapters. It moves the safety interface out to the switching adapter, and eliminates the weight of the power supply XFMR (though you still have the weight of the 2 output XFMRS to deal with).
Last edited:
Those who have designed SMPS for 40 years . . . Probably know more about EMI and EMC,
Versus 90% of those who Either Never Did a SMPS design, Or Never Tried to use a SMPS to power a vacuum tube amplifier.
Versus 90% of those who Either Never Did a SMPS design, Or Never Tried to use a SMPS to power a vacuum tube amplifier.