Fall is upon us and the incessant rain has started in the PacNW, so I am getting back to my Stasis build. The `6 output pairs’ ZM boards are mostly populated and I have turned to the PSU. The intended load is 4 ohms and I am planning on dual mono with CLC 44mF-6mH (0.3 ohms)-44mF using 155B Hammond choke on each rail. Max current of 155B is 2A, but this is within max bias given the limits of the heat sink. Thinking about transformer choice I ran the following estimates:
Dual 400VA AS-4438 -> 38V Secondaries
38*1.414 = 53.7-1.2 (diode loss in rectifier) -> 52.5V
Transformer doesn’t seem to sag much under load (38.3V @5.8A and 37.5V@12.8A), so let's say 52V loaded at idle bias.
Audio power out will be less as output voltage wont swing to rail (Mark Johnson suggests at least 6 Volts from rail, so lets assume 7V) so 52V loaded rail becomes 45V swing
45 /1.414 -> 32V rms
32/4 = 8A rms
32 x 8 = 256W into 4 ohms
However, max bias is limited by heatsinks (4U 400mm) to about 100W.
So 100/45 ->2.2A per side ->185mA per device
2.2 x 2.2 x 4 = 19W Class A
If I do the same for lower voltage secondaries:
30V secondaries but let’s call it ~29.5(loaded)*1.414 = 41.7-1.2 (diode loss in rectifier) -> 40.5V
40.5V loaded rails -~7V as above gives 33.5Vpk
33.5/1.414 = 23.6V rms
23.6 / 4 = 5.9A rms
~140W into 4 ohms
Thermal limitation… 100W -> 100/40.5 = 2.47A per rail (205mA per device)
2.47 x 2.47 x 4 = ~25W Class A
25V secondaries
24.5*1.414 = 34.6V
Estimate @33.4V for loaded rails after rectifier diode loss
-7V gives 26.4Vpk so 26.4/1.414 = 18.6V rms
18.6 / 4 = 4.6A
4.6x4.6x4 = 85W
Max bias = 100/33.4 = 2.99A per side (250mA per device)
3 x 3 x 4 = 36W Class A
So as the available voltage from supply goes down ultimate power goes down, and with it the ability to handle large transients, but power available in Class A goes up because thermal limits at lower rail voltage enables more bias current. I also note that the maximum possible current (loaded rail voltage divided by load of 4 ohms), indicates that 300VA transformer will suffice for 25V secondaries.
Does all this seem approximately right?
Dual 400VA AS-4438 -> 38V Secondaries
38*1.414 = 53.7-1.2 (diode loss in rectifier) -> 52.5V
Transformer doesn’t seem to sag much under load (38.3V @5.8A and 37.5V@12.8A), so let's say 52V loaded at idle bias.
Audio power out will be less as output voltage wont swing to rail (Mark Johnson suggests at least 6 Volts from rail, so lets assume 7V) so 52V loaded rail becomes 45V swing
45 /1.414 -> 32V rms
32/4 = 8A rms
32 x 8 = 256W into 4 ohms
However, max bias is limited by heatsinks (4U 400mm) to about 100W.
So 100/45 ->2.2A per side ->185mA per device
2.2 x 2.2 x 4 = 19W Class A
If I do the same for lower voltage secondaries:
30V secondaries but let’s call it ~29.5(loaded)*1.414 = 41.7-1.2 (diode loss in rectifier) -> 40.5V
40.5V loaded rails -~7V as above gives 33.5Vpk
33.5/1.414 = 23.6V rms
23.6 / 4 = 5.9A rms
~140W into 4 ohms
Thermal limitation… 100W -> 100/40.5 = 2.47A per rail (205mA per device)
2.47 x 2.47 x 4 = ~25W Class A
25V secondaries
24.5*1.414 = 34.6V
Estimate @33.4V for loaded rails after rectifier diode loss
-7V gives 26.4Vpk so 26.4/1.414 = 18.6V rms
18.6 / 4 = 4.6A
4.6x4.6x4 = 85W
Max bias = 100/33.4 = 2.99A per side (250mA per device)
3 x 3 x 4 = 36W Class A
So as the available voltage from supply goes down ultimate power goes down, and with it the ability to handle large transients, but power available in Class A goes up because thermal limits at lower rail voltage enables more bias current. I also note that the maximum possible current (loaded rail voltage divided by load of 4 ohms), indicates that 300VA transformer will suffice for 25V secondaries.
Does all this seem approximately right?
There’s a lot to unpack here. The basic figures are going to be pretty close. There will be variations depending on power transformer size, rectifier Vf and the total bias current running through the output stage.
Dual mono will simplify things as you only need two multimeters per channel while the other is disconnected from power. Probably no need to go for max bias current right away; that will help keep the rail voltages closer to your first approximation.
There is a middle ground between 25V and 30V secondaries. I’m using a pair of Antek AS-3428 transformers. The amp is still a work in progress. It’s a massive beast, intended to drive a pair of large 6 Ohm speakers.
Dual mono will simplify things as you only need two multimeters per channel while the other is disconnected from power. Probably no need to go for max bias current right away; that will help keep the rail voltages closer to your first approximation.
There is a middle ground between 25V and 30V secondaries. I’m using a pair of Antek AS-3428 transformers. The amp is still a work in progress. It’s a massive beast, intended to drive a pair of large 6 Ohm speakers.
@gary s NJW0281/0302
Bias was turned way down to cool off at that moment, trying to determine where it will stabilize. Also PSU only goes to 30V.
Next step will be building/testing 2nd channel, then PSU.
Bias was turned way down to cool off at that moment, trying to determine where it will stabilize. Also PSU only goes to 30V.
Next step will be building/testing 2nd channel, then PSU.
Nice choice for the outputs Jim. What DC rails are you aiming for with the PSU? I am planning on +/- 35VDC rails for mine, that will be plenty of power for me. Bias will be 50degC on heatsinks.
Transformers are 38VAC. Will be interesting to see how close the output can get to the rails. I suspect Vrail -5V or so.
Outputs were mainly chosen because that’s what was available when I ordered. I’ve been sitting on this project for more than a year.
Outputs were mainly chosen because that’s what was available when I ordered. I’ve been sitting on this project for more than a year.
You should get around 51VDC rails or so, depending on mains supply volts and diode bridge Vf drop.
I bought 2 sticks/ea, and didn't do any super fancy matching, just looked at them with a Mega328, sorted as tight as I could, and am putting my trust in that and the 1% 1ohm emitter resistors doing all the heavy lifting when it comes to current sharing.
@6L6 At some point please can you note a safe start up procedure? I omitted to set the position of the trimmers before soldering them on the FE board. Is there a way of powering up, testing and setting the trimmers on ZM's FE boards to a safe value, prior to connecting and applying power to the output stage?
Also, if you get a chance to cast your eyes over the numbers I posted above? Just curious if my conclusion was approximately correct. I am building to power 4 ohm Magnepans...
cheers
Also, if you get a chance to cast your eyes over the numbers I posted above? Just curious if my conclusion was approximately correct. I am building to power 4 ohm Magnepans...
cheers
Don safety glasses, plug it in, and pray the smoke doesn't escape?
The good news is the bias pot is in the 'right' way, and it turns the bias down with counterclockwise rotation. So start with it all the way down. Also thermal equilibrium takes some time to achieve, so adjust slowly.
Yes, your computations look fine. 😎
Well, if by "another one" you mean 'one single amp channel so far', then yes.
🙂 🙂 🙂
The good news is the bias pot is in the 'right' way, and it turns the bias down with counterclockwise rotation. So start with it all the way down. Also thermal equilibrium takes some time to achieve, so adjust slowly.
Yes, your computations look fine. 😎
Well, if by "another one" you mean 'one single amp channel so far', then yes.
🙂 🙂 🙂
As per advice from Jon Soderberg, final adjustments for bias/DC offset should be done with amp chassis all together and with the lid on. He finely tweaks bias slowly up to 50 deg C heat sink temp over 6 hours - that is how they did it back in the Threshold days. Check mV across emitter resistor on each channel for good balance and adjust DC offset to get as close as possible to 0VDC at the output as you go.
Jon also confirmed Hfe matching within 10-20% for N's and P's is fine. Matching N's to P's is not required.
Jon also confirmed Hfe matching within 10-20% for N's and P's is fine. Matching N's to P's is not required.
Last edited:
Man I’ll take anything at this point ha. I’m just happy to see more people building this one. It’s a great amp
Safe start up: place amp at the end of a long extension cord...
Would the MJ/15024/025 be good candidates for the output trannies on a Stasis build? They are rated at 250V/250W with Ft o0f 4Mhz. What were the original outputs?