Power Supply Questions

[rlyach]

I was looking at the power supply schematic, and the transformer shows 160V for the secondary output winding. If I take this as the RMS value with 120V RMS on the primary side, I should get 225V p-p on the secondary side. Then, after the doubler rectifier I get 451V DC. The filter resistors for the 2A3 option are 1200 ohms in parallel with 360 ohms which is 277 ohms. The B+ on the schematic is shown as either 388V or 380V depending if you are looking at power supply output or the channel input. The manual shows 380V. Now, assuming ~54mA of nominal DC current draw, I calculate a B+ voltage of 436V. I must be missing something. I know that I will not get the full 451V DC on the secondary after the doubler, but there is still an discrepancy of ~50V. The only reason I am asking this is that I was going to install a single 1% 270 ohm resistor instead of the two resistors in the kit, and I wanted to calculate what power level resistor I should use.

Thanks

[Grainger49]

Randy,

Are you including the drops through the resistors in the power supply?  Well, my Paramour has one resistor in the power supply.

If you go with one resistor then be certain you have the power rating right.  If you want an upgrade go with a non-inductive wire wound.  The same thing for the cathode resistors.

[rlyach]

The drop in the power supply resistors is already accounted for in my calculation, but the numbers don't add up!?! I suspect that the 160V in the schematic is not correct. If I assume that the B+ in the manual is accurate then the power supply output would be 140V and not 160V. If the 140V is correct, The nominal power dissipation in the power supply resistor is only 0.78 watts. Adding for the full power output to the speakers, a 5W resistor should be sufficient. I was going to use the resistor you mentioned rated at 5W but I wanted to confirm that the 160V in the schematic should really be 140V.

Randy   

[Paul Birkeland]

The simple multiplication misses a lot of variables, and this seems to become more of an issue with very low voltages and voltage doublers in particular.

I'd recommend downloading PSUD from Duncan Amps (it's free), then going through the "edit transformer" dialog and observing the changes to the output voltage based on what you can change for a given transformer that would appear to be the same.

For the Steremour situation, I can get B+ voltages that vary by 50-60 Volts without changing the secondary voltage.

[Paul Joppa]

Real-world power supplies with cap-input filters usually produce around 1.2 times the rms input; it's only when they are unloaded that they produce close to 1.41 times. The biggest cause is the resistance of the windings. In Stereomour, the primary is about 3.5 ohms and the HV secondary about 10 ohms.

PB's suggestion of PSUD is right on the money - it's very educational to play around with it. I did a quick model with the above parameters and got 386 volts.

(In addition to correcting the primary resistance for the stepup ratio squared, technically you must also subtract the primary voltage lost by its resistance due to the filament and heater power. Fortunately this is small, about a volt off the HV secondary.)

[rlyach]

Thanks for putting up with a noob. I completely ignored the output resistance of the transformer. I tried PSD and could not duplicate your numbers. I looked up the ESR of the filter caps (panasonic I think) and used 277 ohm for the filter resistor. I used a voltage doubler topology. If I use 50 ohms for the transformer output resistance I get 383V. If I use 10 ohms I get 415 V. Not sure what I am doing wrong. Here are my results with the model.

[Paul Joppa]

It's a stereo amplifier  :^) 

You have modeled one channel with a 7K resistor drawing about 55mA. To get my numbers, you have to use a 135 ohm resistor (two 270's in parallel) and a 3500 ohm load resistor drawing 110mA.

BTW, I've made the same mistake myself, more than once.

[rlyach]

Duh... That finally occurred to me. I use 139 ohms and 3500 ohms I get 403 volts. Still too high. BTW I also used a 200uF cap with half the ESR across the load.

[rlyach]

The bottom line is... If I want to replace the resistors in the kit with better resistors I should use 270ohm 10W (5 W is pretty close) and 1200 ohms for the cathode bias at 10W. I want Ohmite audio gold resistors but they are very hard to find. I think I will stick with the kit resistors. Non-inductive 1% power resistors are too hard to find.

[Jim R.]

You can always buy Mills MR-12 types and ask a place like partsconnexion to match them for you.  I do it all the time.

-- Jim

[Grainger49]

Here is a link to the Mils main page:

http://www.partsconnexion.com/resis_mills_main.html

[rlyach]

Thank you so much. I really wanted to use just one resistor to match the channels more effectively. I searched for non-inductive wire wound resistors and even called my local electronics supply place. For some reason the Mils resistors never showed up in anyone's searches. I just ordered the resistors MR-12, 270 ohm and 1200 ohm. Thanks again.

[Jim R.]

Randy,

My pleasure -- also, they happen to be fine sounding resistors, or maybe I should say they don't do anything to muck up the sound.  They work great in tube amp power supplies and as cathode resistors.

-- Jim

[rlyach]

I really want to thank everyone for putting up with my questions. I am really trying to learn as much as I can about this amp before I build it. I found an on-line circuit simulator (http://www.circuitlab.com/) that allows any circuit construction, almost all common elements (no tubes), and the ability to edit parameters like ESR. Additionally, you can run DC, time domain, and frequency domain simulations. The system works very very well. I put the Stereomour power supply in the system and it output 387 volts using resistors, and 383 volts using choke loads.

Randy

[rlyach]

My kit should arrive on Monday so I have had time for one more simulation. I put the entire schematic in the circuit lab simulator. I modeled the tubes as voltage controlled voltage sources with offsets to account for the bias points. The simulation results agree very well with the documentation. This only shows that the simulation is more or less correct; we already know the design is. The constant current source outputs just under 4 mA. The 2A3 bias current is 50mA. It took the model 0.5s for the circuit to stabilize. I used a 2V p-p 1KHz signal for the input. As you can see from the waveforms, The grid voltage on the 2A3 swings from -58V to +58V. I did not show it, but the plate voltage on the 12AT7 swings from 136V to 252V. You can also see the phase shift of 180 degrees between the first and second amplifier stages. The plate of the 2A3 swings from just above 500V down to a little below 200V. On the other side of the parafeed cap the signal is shifted to be centered  around 60V. The output of the transformer is +2.24 to - 2.23 into an 8 ohm resistive load. It was very educational to see all this working together.