Imho this is diy stuff. There are likely few who have used this smps module, and fewer who have made some effort to check and compare performance with different configurations and how they relate to more difficult to appreciate concerns like ground related hum and noise and hf ripple. Perhaps unless you can identify someone who has written up their project, and you can follow it pedantically to make your project, then you may be out of your depth unless you can learn from your own diy experiences.
not even sure if the 12.6V@4A and the 6.3V@4A are two separate supplies, since i can not examine the psu in person.....
so it is really up to you and the tubes you are using in your actual application, we can not guess in front of a pc monitor......
otherwise, you can connect two 6.3v filaments to use the 12V only if they have the same current draw. otherwise there is some balancing acts needed to make sure that the voltages are 6.3 across filament pins...
so it is really up to you and the tubes you are using in your actual application, we can not guess in front of a pc monitor......
otherwise, you can connect two 6.3v filaments to use the 12V only if they have the same current draw. otherwise there is some balancing acts needed to make sure that the voltages are 6.3 across filament pins...
As I said earlier I bought one of these some years ago but haven't tried it out yet. I have a different project breadboarded at the moment (a direct coupled SET with all DH tubes) so it will be a while before I can mess with this.
So I have no schematic yet, just a general plan of using it to power a very simple SE amp - two output tubes and no more than two input tubes. Outputs will likely be something like 6BQ5, 6V6, 6AQ5 or possibly 6N6G.
Earlier in this thread there were a few posters who seemed to indicate that they were using this SMPS either experimentally or in a build. I referred to some of their comments in my posts above.
Starting in Post #11 @allevmail says: "I just tried this PSU in my mini 6V6 SE amp (2 x 6V6 + 6SN7) . Had some problem with HV delayed start so reduced delay to 3-5 seconds by changing resistor on NE555. Now it works OK. No audible noises. No overheating (no fan required). . . . It does power up the heaters to a stable condition all the time, no problem. The problem was just inrush of high voltage . . . After delay disabled all works OK." Unfortunately, he hasn't been on DIY since Dec 2021.
@kodabmx , after posting his frustrations that it wouldn't work for an 8 tube preamp with a large cap later said: "I used this SMPS for a 6F12P/6P43P Little Miracle amp (6 tubes total) and I get 12.6V and 292V from it. I'm "burning" it in right now.
My trick was to tack a 33k resistor across the timing resistor for the 555 so the HV only delays 1/2 seconds instead of 30. There is no capacitor on the HV for output tubes, and RC of 10k/50uF per channel for the VA/PI stage. It hummed like a bastard on the bench, but it's quiet in the rack."
@Tubekwk describes some testing he did in Post #72 and later posts: "The power supply has been running for 35 minutes with the fan on and the heatsink and flyback transformer are only just warm, a good result indeed.
6.3V load 2 x EL34s and one 12AX7
300V load 1000 ohms."
In post #91 @jcalvarez says: "I also was explained that I need to load the 12V output to get the maximum HV voltage. I got around 295, I can live with that. I use mainly for experiments, seems to be relatively quiet."
This is why I asked about how to run both the 6.3v and 12.6v heaters, although earlier posters seem to be getting 300v output without a load on the 12.6v output. Hence my confusion.
And then recently he says "I'd recommend just using the smps as is. Connect the output tubes in series to the 12.6V output, and the input tubes to the 6.3V or 12.6 depending on the heater requirements. Soft start is not reay needed for such low voltage, and you already have an HV delay."
This seems to indicate two things . . . 1) that the 30 second HV delay does not need to be disabled or modded??? and 2) that both heater outputs can be run at the same time. So I asked how to wire them since they share a single ground.
At this point I'm just trying to find out:
A) does the HV delay need to be modded and, if so, perhaps someone could show me which resistor needs to either be replaced or paralleled with a 33k.
B) answers to the questions I posted above in Post #114, including how to "isolate" the HV and heater grounds from each other. I thought my questions were pretty clearly worded and straightforward, but if they're confusing please say so and I'll try my best to clarify them.
do not expect to accomplish anything by merely asking, do it and learn, you can be a lot happier that way.....
@FlaCharlie I'm travelling right now, I should be able to draw something in a couple of days.
But it all comes to this:
You can use B+ directly for the output stage. Add an RC filter for the input stage.
About heaters: the reason they have separate ground for B+ and heaters is to allow heater elevation if required. You can wire separately the heaters and B+ circuits, then join the two grounds near the power supply.
How to wire up the heaters: Lets say you use two 6V6 for output and a 12AX7 for input stage. Connect the 6V6 heaters in series, then wire them to the 12.6V smps output. For tbe 12AX7, two options a) pin 4&5 shorted, and apply 6.3V to them, pin 9 to the ground, or b) just connect pint 4 to 12.6V and pin5 to ground, leave pin 9 unconnected.
But it all comes to this:
You can use B+ directly for the output stage. Add an RC filter for the input stage.
About heaters: the reason they have separate ground for B+ and heaters is to allow heater elevation if required. You can wire separately the heaters and B+ circuits, then join the two grounds near the power supply.
How to wire up the heaters: Lets say you use two 6V6 for output and a 12AX7 for input stage. Connect the 6V6 heaters in series, then wire them to the 12.6V smps output. For tbe 12AX7, two options a) pin 4&5 shorted, and apply 6.3V to them, pin 9 to the ground, or b) just connect pint 4 to 12.6V and pin5 to ground, leave pin 9 unconnected.
@jcalvarez Thanks for your helpful reply.
It seems that you didn't find it necessary to mod or disable the 30 second delay on the HV. Is that correct?
I wonder if the problems with the delayed start are linked to the HV current draw / wattage demands of the circuit.
This seems to be supported by @Tubekwk 's comment about his initial testing. He said "Turned on, the power supply failed to start (hiccup mode) so I turned down the output voltage pot and the device started at 260VDC output, I slowly turned the pot up to maximum and 301VDC was measured." For this initial test he was using "a 500 ohm 200W load on the 300VDC output." So he was trying to see how it reacted to a 600mA load, which is the advertised maximum.
His next test was "Two EL34s and one 12AX7 were connected to the 6.3VDC output, the cold resistance of this circuit was 0.3 ohms. A single 1000 ohm 100W resistor was used to load the 300VDC supply to 300mA. The power supply started and was left to run for 1 hour."
I'm guessing that my total current draw will be in the range of 100 to 125mA. Perhaps those who found it necessary to mod or disable the 30 second delay were drawing more current and/or also adding an RC section to supply the B+ using a cap value that might also create issues.
If others with hands on experience with this SMPS care to offer any practical tips or supply details of their implementation I would certainly appreciate it.
It seems that you didn't find it necessary to mod or disable the 30 second delay on the HV. Is that correct?
I wonder if the problems with the delayed start are linked to the HV current draw / wattage demands of the circuit.
This seems to be supported by @Tubekwk 's comment about his initial testing. He said "Turned on, the power supply failed to start (hiccup mode) so I turned down the output voltage pot and the device started at 260VDC output, I slowly turned the pot up to maximum and 301VDC was measured." For this initial test he was using "a 500 ohm 200W load on the 300VDC output." So he was trying to see how it reacted to a 600mA load, which is the advertised maximum.
His next test was "Two EL34s and one 12AX7 were connected to the 6.3VDC output, the cold resistance of this circuit was 0.3 ohms. A single 1000 ohm 100W resistor was used to load the 300VDC supply to 300mA. The power supply started and was left to run for 1 hour."
I'm guessing that my total current draw will be in the range of 100 to 125mA. Perhaps those who found it necessary to mod or disable the 30 second delay were drawing more current and/or also adding an RC section to supply the B+ using a cap value that might also create issues.
If others with hands on experience with this SMPS care to offer any practical tips or supply details of their implementation I would certainly appreciate it.
Correct, I did not disable it.
I did not try to load the smps with the vendor's maximum current. My experience with Aliexpress tells me that 50% is a safe bet. In my tests I drove it up to 240mA, the transformer was warm, but thete was no need for a fan.
If your B+ load current is about 150mA then you will have no problems.
If you want to learn about the 555 timer and where the resistor is - https://www.learningaboutelectronics.com/Articles/555-timer-delay-before-turn-on-circuit.php
The 555 is an 8-leg IC, it looks like there are 2 IC's on the board with 8 legs so one of them has to be the 555.
Thanks. I found it, although it seems like I won't need to mod it since I won't be drawing much current.
The 555 is the chip behind the power outputs... The timing resistor is on pin 6.
I disabled the timer because it would heat the tubes for 30 seconds and drop out when B+ came on, but by the time 30 seconds goes by, the PSU caps are discharged again and the cycle repeats.
I'm using a 100R 47uF filter on the HV.
The delay is now about 1/2 a second. It restarts several times before it latches (when the tubes are warmed a bit, and the main cap is charged).
The amp will turn itself off with a good bass hit at full power, too... And it doesn't restart until the tubes are cold enough to not conduct anymore so the HV can latch up again...
BUT I use no fan or ventilation and it works at 75 degrees for days at a time 🙂
I disabled the timer because it would heat the tubes for 30 seconds and drop out when B+ came on, but by the time 30 seconds goes by, the PSU caps are discharged again and the cycle repeats.
I'm using a 100R 47uF filter on the HV.
The delay is now about 1/2 a second. It restarts several times before it latches (when the tubes are warmed a bit, and the main cap is charged).
The amp will turn itself off with a good bass hit at full power, too... And it doesn't restart until the tubes are cold enough to not conduct anymore so the HV can latch up again...
BUT I use no fan or ventilation and it works at 75 degrees for days at a time 🙂
So I'm using this SMPS on the breadboard running a simple SE circuit. First version used 6N6G output tubes and a single 6SN7 input tube. Later I switched to 6V6 outputs using the same input tube. I'm using the unit in stock form, no mods to the HV delay. I'm running the B+ straight off the 300v and then using a 1.8k resistor and 22uf cap to supply the 6SN7 plates. I just hooked it up and it worked immediately without issue.
High voltage comes up without any hiccups with less than a 30 sec delay. There's a very subdued "thump" when the HV comes on but I'm not concerned about that.
I'm running the heaters in parallel off the 6.3v and they are floating, the 0v heater connection which is shared with the 12.6v only connects to the tube pins, not to ground.
Unless I misunderstood, earlier someone said that the 12.6v heater connection had to be loaded for this SMPS to function. I have not found that to be the case. There is no connection to the 12.6v.
In fact, I'm having problems whenever I use the 12.6v connection, which is what I'd like to ask about.
Earlier I was using 6N6G output tubes instead of 6V6s and I decided I would try using a 12SN7 in place of the 6SN7. So I wired it with the 12.6v to one pin and the shared 0v to the other pin. Floating, no ground, same as the all 6.3v setup.
When I did this it would go into hiccup mode. I tried grounding the 0v connection but the results were the same. After changing the outputs to 6V6s, which draw less heater current than the 6N6Gs, the results were the same. I then tried running the 12SN7 with 12V6 outputs. Heaters wired exactly like the all 6.3v version which works fine . . . heaters floating with nothing connected to the 6.3v. Still goes into hiccup mode.
Excessive heater current draw doesn't seem to be the issue. The combo of 6SN7 and two 6N6Gs draws a total of 2.2A and the 6SN7 / 6V6 combo draws 1.5A. By contrast, the 12SN7 / 12V6 combo draws the least current at 0.75A.
I would like to have the option of running the 12SN7 / 12V6 combo since the tubes are so cheap. Any idea how to make the 12.6v output behave like the 6.3v one does?
High voltage comes up without any hiccups with less than a 30 sec delay. There's a very subdued "thump" when the HV comes on but I'm not concerned about that.
I'm running the heaters in parallel off the 6.3v and they are floating, the 0v heater connection which is shared with the 12.6v only connects to the tube pins, not to ground.
Unless I misunderstood, earlier someone said that the 12.6v heater connection had to be loaded for this SMPS to function. I have not found that to be the case. There is no connection to the 12.6v.
In fact, I'm having problems whenever I use the 12.6v connection, which is what I'd like to ask about.
Earlier I was using 6N6G output tubes instead of 6V6s and I decided I would try using a 12SN7 in place of the 6SN7. So I wired it with the 12.6v to one pin and the shared 0v to the other pin. Floating, no ground, same as the all 6.3v setup.
When I did this it would go into hiccup mode. I tried grounding the 0v connection but the results were the same. After changing the outputs to 6V6s, which draw less heater current than the 6N6Gs, the results were the same. I then tried running the 12SN7 with 12V6 outputs. Heaters wired exactly like the all 6.3v version which works fine . . . heaters floating with nothing connected to the 6.3v. Still goes into hiccup mode.
Excessive heater current draw doesn't seem to be the issue. The combo of 6SN7 and two 6N6Gs draws a total of 2.2A and the 6SN7 / 6V6 combo draws 1.5A. By contrast, the 12SN7 / 12V6 combo draws the least current at 0.75A.
I would like to have the option of running the 12SN7 / 12V6 combo since the tubes are so cheap. Any idea how to make the 12.6v output behave like the 6.3v one does?
The header SMPS needs twice the rating of the heater current. When cold they draw a lot more current.
The 6.3V is derived from the 12.6V output (or at least that is claimed in Aliexpress), therefore loading it will load the 12.6V. And it is not strictly required, but the HV will be lower if you do not load the heater output. I definitely saw that, reported to the seller that the HV was too low, and he explained that I needed to load the heater output for the HV to give the expected voltage.
The 12.6V output is rated at 4A, it should easily handle 0.75A cold heater.
This is strange, when I tested mine I used a 20W 12V car bulb, and it powered up without hiccup. I will un-dust mine a bit later and try, and while I do not have a high current 12V tube, a couple of 6P45S in series should do.
That 0,75A should be hot nominal current , cold is much higher and tube filament is slowly warming not like a bulb that is instant hot and glowing.
For a power supply with current/overpower protection it is very tricky
For a power supply with current/overpower protection it is very tricky
Not that difficult, I just had a similar situation with a small smps (15V 25W) powering 4 PCL84s. It would take 30s to stop "hiccupping". A simple circuit solved the problem, now it is just two 2~3s and the output voltage stabilises.
Last edited:
Looks like I figured out a few things though I'm not sure why it acts as it does.
For one, I discovered that it's best to ground the 0v terminal that's shared by the heater outputs. Doing so doesn't seem to affect the heaters but with the heaters floating the B+ jumped to ~365v very briefly at startup. This shouldn't be an issue since the caps are rated at 400v but with the heaters referenced to ground there was no voltage spike on startup so I left them grounded.
With the 6SN7 / 6V6 combo installed I measured the unloaded 12.6v connection at ~13.35v which wasn't a surprise. The single voltage adjustment pot, which affects all voltages, had been set so that I had 6.3v on the heaters which gave me a B+ of ~295v.
I left the 6V6s in and subbed a 12SN7 for the 6SN7. On the chance that it might be shutting down because the voltage on the 12.6v tap was too high, I turned down the voltage adjustment pot a good bit and after doing so it started fine. So I adjusted the pot higher but it still hiccuped so I repeated the process . . . shut down, let the tubes cool, and adjusted the pot a tiny bit lower. I repeated this process until the hiccuping stopped. I neglected to take voltage measurements at this point.
Next I subbed in 12V6s so now all tubes are on the 12.6v and nothing is connected to the 6.3v. On the first attempt it shut down so I adjusted the pot a tiny bit lower. Repeated the process until the hiccups finally stopped, which took quite a bit of adjustment.
At that point the voltages were only 11.68v on the heaters and 267.8v B+.
So it seems that the more that's connected to the 12.6v section the less B+ you get.
The next step was to put the 6N6G output tubes back in, leaving the 12SN7 in place. The voltage pot was still set for the 12SN7 / 12V6 combo so it started right up. I then adjusted the voltage up bit by bit until I found the highest setting that didn't hiccup.
At that point I measured 12.42v, 6.03v and 286.2v B+.
Then I went back to using a 6SN7 with the 6N6Gs. The voltage pot was still set for the 12SN7 / 6N6G combo so it started right up. I then adjusted the voltage up so that I got 6.3v on the heaters. Now the 12.6v is unloaded. Turned it off, let tubes cool and it starts up at that setting without issue.
I now have 13.22v on the unloaded 12.6v output, 6.3v and a B+ of 299.8v.
I understand the need to derate the current and that the filaments draw considerably more current when cold. But the 12SN7 / 12V6 combination have a total current draw 0.75A which means they draw 12.6v x 0.75A = 9.45w. On the other hand the 6.3v combo of the 6SN7 and the 6N6Gs have a total current draw of 2.2A. So 6.3v x 2.2A = 13.86w, which is 46.6% more heater wattage yet the B+ is higher.
I didn't measure the 12V6s but they are biased the same as the 6V6s which were drawing ~44mA each while the 6N6Gs draw 46.5mA each so I would expect B+ to be lower not almost 12% higher.
So the highest (no hiccup) voltage with the 12SN7 / 12V6 combo was only 11.68v on the heaters and 267.8v B+ even thought the total current draw is lower. And the 6SN7 / 6N6G combo can run at 6.3v with a B+ of 299.8v even though the total current draw is higher. It might even go higher without hiccuping but I don't want to run the heaters over 6.3v.
For one, I discovered that it's best to ground the 0v terminal that's shared by the heater outputs. Doing so doesn't seem to affect the heaters but with the heaters floating the B+ jumped to ~365v very briefly at startup. This shouldn't be an issue since the caps are rated at 400v but with the heaters referenced to ground there was no voltage spike on startup so I left them grounded.
With the 6SN7 / 6V6 combo installed I measured the unloaded 12.6v connection at ~13.35v which wasn't a surprise. The single voltage adjustment pot, which affects all voltages, had been set so that I had 6.3v on the heaters which gave me a B+ of ~295v.
I left the 6V6s in and subbed a 12SN7 for the 6SN7. On the chance that it might be shutting down because the voltage on the 12.6v tap was too high, I turned down the voltage adjustment pot a good bit and after doing so it started fine. So I adjusted the pot higher but it still hiccuped so I repeated the process . . . shut down, let the tubes cool, and adjusted the pot a tiny bit lower. I repeated this process until the hiccuping stopped. I neglected to take voltage measurements at this point.
Next I subbed in 12V6s so now all tubes are on the 12.6v and nothing is connected to the 6.3v. On the first attempt it shut down so I adjusted the pot a tiny bit lower. Repeated the process until the hiccups finally stopped, which took quite a bit of adjustment.
At that point the voltages were only 11.68v on the heaters and 267.8v B+.
So it seems that the more that's connected to the 12.6v section the less B+ you get.
The next step was to put the 6N6G output tubes back in, leaving the 12SN7 in place. The voltage pot was still set for the 12SN7 / 12V6 combo so it started right up. I then adjusted the voltage up bit by bit until I found the highest setting that didn't hiccup.
At that point I measured 12.42v, 6.03v and 286.2v B+.
Then I went back to using a 6SN7 with the 6N6Gs. The voltage pot was still set for the 12SN7 / 6N6G combo so it started right up. I then adjusted the voltage up so that I got 6.3v on the heaters. Now the 12.6v is unloaded. Turned it off, let tubes cool and it starts up at that setting without issue.
I now have 13.22v on the unloaded 12.6v output, 6.3v and a B+ of 299.8v.
I understand the need to derate the current and that the filaments draw considerably more current when cold. But the 12SN7 / 12V6 combination have a total current draw 0.75A which means they draw 12.6v x 0.75A = 9.45w. On the other hand the 6.3v combo of the 6SN7 and the 6N6Gs have a total current draw of 2.2A. So 6.3v x 2.2A = 13.86w, which is 46.6% more heater wattage yet the B+ is higher.
I didn't measure the 12V6s but they are biased the same as the 6V6s which were drawing ~44mA each while the 6N6Gs draw 46.5mA each so I would expect B+ to be lower not almost 12% higher.
So the highest (no hiccup) voltage with the 12SN7 / 12V6 combo was only 11.68v on the heaters and 267.8v B+ even thought the total current draw is lower. And the 6SN7 / 6N6G combo can run at 6.3v with a B+ of 299.8v even though the total current draw is higher. It might even go higher without hiccuping but I don't want to run the heaters over 6.3v.
Last edited:
I'm listening to a Little Miracle amp that I power with one of these.
I had the issue of all the tubes warm up but when it turns on B+ the supply turns off and won't restart until the tubes are cold enough to stop conducting.
So I reduced the 30 second delay to about 300ms, and let it "hiccup" roughly 10 times on start - that's long enough to charge the B+ caps and get the heaters warm enough to let it latch up.
If you drive it to clipping and a bass drum hits, it'll shut off though 🙂
Like most no name SMPS, the power rating is for resistive loads, and for practical purposes it should be derated to about 40% if you run it in a closed chassis like I do.
I had the issue of all the tubes warm up but when it turns on B+ the supply turns off and won't restart until the tubes are cold enough to stop conducting.
So I reduced the 30 second delay to about 300ms, and let it "hiccup" roughly 10 times on start - that's long enough to charge the B+ caps and get the heaters warm enough to let it latch up.
If you drive it to clipping and a bass drum hits, it'll shut off though 🙂
Like most no name SMPS, the power rating is for resistive loads, and for practical purposes it should be derated to about 40% if you run it in a closed chassis like I do.
I'm not getting any hiccups at all with the two SE setups that use all 6.3v tubes. As noted, with the other tube combos which include at least one 12.6v tube, lowering the voltage eliminates the hiccups but I'm not able to run the heaters at full voltage like I can with the 6.3v tubes and the B+ is lower.
The SMPS is stock and the startup delay is only ~17 sec. Once the B+ comes on it stays on, never had it shut down while playing.
I run B+ directly off the 300v output then a 1.8k R and 22uf cap to supply the 6SN7 plates. It's dead quiet, no hum or noise at all.
The circuit doesn't draw much current though - only ~103mA total for the 6SN7 / 6N6G version and ~98mA with the 6V6s.
How large is your B+ cap? You might try reducing its value or even eliminating it entirely. Since my setup is dead quiet without any cap I assume the supply itself is well filtered. My meter shows no AC voltage at all on the 300v output and playing bass heavy music hasn't caused any issues.
Last edited: