Hey everyone,
i want to build a PP DHT amp next and was thinking about using the 811a as they can work on relatively low B+ and are still affordable (at least the russian and chinese versions). I came up with the following schematic:
Spice says 80W at 1.8% THD but the 811s are at running hard at that point drawing 350mA peak at the plates and 22mA peak at the grids. May have to dial it back a bit. Idle current for the 811s is set to 28mA.
For heating the 811's i'd use 2x MeanWell 3.3V supplies in series. I can connect all output tubes to the same (floating) smps when using fixed bias correct?
Now i still have a few questions..
Looking forward to your input!
i want to build a PP DHT amp next and was thinking about using the 811a as they can work on relatively low B+ and are still affordable (at least the russian and chinese versions). I came up with the following schematic:
Spice says 80W at 1.8% THD but the 811s are at running hard at that point drawing 350mA peak at the plates and 22mA peak at the grids. May have to dial it back a bit. Idle current for the 811s is set to 28mA.
For heating the 811's i'd use 2x MeanWell 3.3V supplies in series. I can connect all output tubes to the same (floating) smps when using fixed bias correct?
Now i still have a few questions..
- Are there any obvious mistakes in my schematic?
- Is the Hammond 1650R a suitable output transformer? I suppose it would saturate at max. output, not like i'd ever run it at full power tho.
- Should i regulate the +/- 100V Supply for the mosfets? Don't want the supply to sag too much when the grids draw current.
- Are the russian G811's any good? Alternative would be the Psvane ones.
Looking forward to your input!
That should work OK as drawn.
If you are using a pair of 3.3 volt DC supplies, you would tie all of the 811 filaments together and ground one side as long as you are biasing with the mosfets. Note that is 6.3 v 16A required. Remember the 811a filaments will have a lower resistance when cold and you should verify that the supply will start OK. Some SMPS supplies can have trouble starting.
The 1650R output transformer is a good choice for 811a push-pull. I have run the 1650R at about 800 volts on the plates just fine though that was with a powered subwoofer so from about 65 hz and up.
I have used mosfet drive on 811As. Grid on resistance will be about 1K depending on plate voltage. You could drop the grid stopper to 50 ohms.
You could choose a smaller driver mosfet like something with a 40W rating to get lower reverse gate capacitance.
The mosfet follower has a fairly high psrr so the +- 100 volt supply is not so critical. You need something like an 80 volt winding if using half wave rectification or 160 VCT for full wave.
Have a look at the Altec 1570B for a good example of an amplifier using 811A tubes also.
If you are using a pair of 3.3 volt DC supplies, you would tie all of the 811 filaments together and ground one side as long as you are biasing with the mosfets. Note that is 6.3 v 16A required. Remember the 811a filaments will have a lower resistance when cold and you should verify that the supply will start OK. Some SMPS supplies can have trouble starting.
The 1650R output transformer is a good choice for 811a push-pull. I have run the 1650R at about 800 volts on the plates just fine though that was with a powered subwoofer so from about 65 hz and up.
I have used mosfet drive on 811As. Grid on resistance will be about 1K depending on plate voltage. You could drop the grid stopper to 50 ohms.
You could choose a smaller driver mosfet like something with a 40W rating to get lower reverse gate capacitance.
The mosfet follower has a fairly high psrr so the +- 100 volt supply is not so critical. You need something like an 80 volt winding if using half wave rectification or 160 VCT for full wave.
Have a look at the Altec 1570B for a good example of an amplifier using 811A tubes also.
Thanks, good to know. I'll overspec the smps quite a bit, meanwell has 3.3V/30A supplies available that can be adjusted down to 3.14v. I'll build a separate enclosure for the psu anyways so i'll have plenty of space for big smps's.
I'll drop the grid stopper value. a quick search for available mosfets turned up the STP 10NK60 which has 35W dissipation, it has the protective zeners integrated as well so i could omit those.
Not sure which transformer i can get for the mosfet supply yet.
What's the rule of thumb for PT current rating in AB amps? Since it only draws 700mA peak with both channels driven to max power i'd assume you don't need a winding that can provide a lot more than that, maybe 800mA would be fine? Either way looks like i'll have to have a custom transformer wound.
I'll drop the grid stopper value. a quick search for available mosfets turned up the STP 10NK60 which has 35W dissipation, it has the protective zeners integrated as well so i could omit those.
Not sure which transformer i can get for the mosfet supply yet.
What's the rule of thumb for PT current rating in AB amps? Since it only draws 700mA peak with both channels driven to max power i'd assume you don't need a winding that can provide a lot more than that, maybe 800mA would be fine? Either way looks like i'll have to have a custom transformer wound.
Hammond may have something that will work for you for both plate and bias.
About 500-600mA may be reasonable unless you really plan to drive both channels continuously.
https://www.hammfg.com/electronics/transformers/classic
https://www.hammfg.com/electronics/transformers/classic/300
https://www.hammfg.com/electronics/transformers/classic/261-262
About 500-600mA may be reasonable unless you really plan to drive both channels continuously.
https://www.hammfg.com/electronics/transformers/classic
https://www.hammfg.com/electronics/transformers/classic/300
https://www.hammfg.com/electronics/transformers/classic/261-262
With Mosfet drivers and positive bias you might want to take a good look at the startup sequence; if the positive grid voltage comes up before the plate voltage the grids will draw a lot of current. Probably only a problem if you're using IDH rectifiers for the HT.
And yes, the SMPS filament supplies must be rated for much more than 4A / tube, the cold resistance is close to a dead short. You could use 9V power supplies and add 0,68R in series with each filament to get a nice soft start with a much lower inrush current, though each resistor will produce almost 11 watts of heat (and much more for the first few seconds).
And yes, the SMPS filament supplies must be rated for much more than 4A / tube, the cold resistance is close to a dead short. You could use 9V power supplies and add 0,68R in series with each filament to get a nice soft start with a much lower inrush current, though each resistor will produce almost 11 watts of heat (and much more for the first few seconds).
If you use the same filament supply for both tubes you will be shorting their "cathodes" together making individual current measurements impossible in the cathode circuit since the 811A is a DHT. The TSE and TSE-II amps get around this by putting the current sense resistor in the B+ lead of the OPT, but that won't work here since this is a push pull amp.
I would use a 7.5 volt supply for each tube with a dropping resistor.
As daqvin_carter stated the power transformer does not need to be rated for the "full power" current of both channels unless you intend to run continuous sine waves for long periods of time. This is a HiFi amp not a shaker table driver. Most HiFi amps will be making less than 10% of their full power rating when playing music that hits clipping on peaks.
I would use a 7.5 volt supply for each tube with a dropping resistor.
As daqvin_carter stated the power transformer does not need to be rated for the "full power" current of both channels unless you intend to run continuous sine waves for long periods of time. This is a HiFi amp not a shaker table driver. Most HiFi amps will be making less than 10% of their full power rating when playing music that hits clipping on peaks.
7.5V would be more reasonable than 9V, but a bit harder to find. I had to dig deep into eBay when I bought my 7.5V 19A switch mode power supplies years ago as a makeshift solution for heating up a pair of 808s, with the option to add 0R3 dropping resistors and use 811A once I ran out of 808s.
Back in the late 1990s I found some 5 volt switch mode supplies that could be turned up to 6.3 volts as long as you removed the output crowbar. You can also use cathode feedback around the output side of the transformer if they have the right windings. I found it was be best configuration overall. Not sure of the current Hammond 1650R allows that or not.
Then kids happened...
Now all kids can drive themselves and have jobs. I will soon have a man cave devoted to high efficiency speakers and the main living room will have the lower efficiency speakers using mostly Eton drivers. Power amplifier will likely have a 1650R transformer driven by 4x150s instead of 811As.
Then kids happened...
Now all kids can drive themselves and have jobs. I will soon have a man cave devoted to high efficiency speakers and the main living room will have the lower efficiency speakers using mostly Eton drivers. Power amplifier will likely have a 1650R transformer driven by 4x150s instead of 811As.
Last edited:
For those in the USA, Jameco has a large selection of SMPS's. I got some open frame 7.5 volt units a couple years ago to go in a modular music synthesizer. I ran 7.5 volts to every row of modules with a 5 volt LDO at each row. Ditto some 15 volt units to feed 12 volt LDOs.
A quick look at the Jameco web site finds 46 different 7.5 volt SMPS's but only 4 are actually in stock. Many of the non stocked units only have a 2 or 3 week lead time. One of the stocked units is a 7.5 volt 8 amp open frame MeanWell for $15.40.
The 5 volt generic SMPS I got from Amazon will only go to 5.65 volts. No, OVP trip, just runs into the end of the pot. This was not good enough to make the LDO happy.
My power supplies are chinese, MeanWell or similar brand of unknown quality. The high shipping costs makes it more or less impossible to buy anything from USA, a 9 pin tube can cost 20-30 dollars to get here🤐
I still haven't listened to an SMPS-heated DHT but the power supplies are at least good enough to run some measurements on the tube.
After reading the long thread about Colemans filament regulators, SMPS seems like one of the worst options for audiophile-approved DHT heating.
I still haven't listened to an SMPS-heated DHT but the power supplies are at least good enough to run some measurements on the tube.
After reading the long thread about Colemans filament regulators, SMPS seems like one of the worst options for audiophile-approved DHT heating.
Hmm, many of the good ones are 115v primary only, being in germany i need 230v. Will have a closer look tomorrow but i think a custom one will be the most economic option as well (i paid just 80€ for a custom toroid with 650 and 250v windings for my gm70 amp).
Fuling, good thing you mention that. i would have turned on the heaters + bias first and had the B+ and mosfet supply on a time delay. So now i will do heaters only at first and then switch on B+, Mosfet supply and Bias supply at the same time.
Tubelab, you're exactly right. i completely forgot about measuring the cathode current on each tube for setting the bias, i will use one psu per tube. I see meanwell has a 7.5V/8A one available for just 20€, may try that.
Not sure about the sonic differences between smps and linear supply on heaters tbh, sure i can imagine if it's a really low quality smps with tons of switching noise that could be an issue. On my gm70 amp i use 180W 19.5V Dell laptop adapters. i may take measurements to see if there is any ripple/switching noise, just waiting for my output transformers still.
One tactic may to just have a relay the shorts the mosfet gates to ground until there is a little warmup and then open.
For Class B or AB2 the loadline usually needs to be adjusted to a higher value.
Otherwise the plates run straight into the diode line of the tube. The diode line is plate current saturation. 👍
Otherwise the plates run straight into the diode line of the tube. The diode line is plate current saturation. 👍
Years ago I took some regular 5v switching supplies and added a bit of resistance to the trimmer to bring the voltage up to 6.3.
When I built that PP 6V6 Class AB2 amp I loaded 6.4K P-P rather than leaving it at 8K as in the Class AB1 mode.
Best bet is have a look at the plate family of curves. In post 13, think I said higher value, should be lower.
Luckily no mistakes with the chain saw today tho. 👍
😀
Best bet is have a look at the plate family of curves. In post 13, think I said higher value, should be lower.
Luckily no mistakes with the chain saw today tho. 👍
😀