PSUD2 Models for popular amps

[dbishopbliss]

I'm trying to find the values to use in the power supply for the Supro 1690T amp I am building. The values specified on the schematic I am using are raising errors in PSUD2. I think I must be entering something wrong so I would like to compare the values with models of other amps. Problem is, I cannot find any examples of popular amps.


Has anyone modeled a 5E1 Champ or a 5E3 Deluxe or something common like that?

[sluckey]

I fear you have lost focus and fallen down a deep rabbit hole. Call Alice. I think she'll know. 

[PRR]

What "errors" are you seeing??

What values are you using? Can you screen-shot PSUD?

> the schematic I am using

Show that also.

[dbishopbliss]

Before I get to my errors, here is what I'm doing for a simple model of the 5E3 5E1 Champ.


The rectifier is a Full-Wave Tube style. I changed the rectifier to 5Y3-GT and set the transformer voltage to 325V. I don't know the Source resistance so I just left the value that was there.


Then I added an LC filter and an RC filter using the values for the champ. I had to look up the specs on the choke and 4H 167ohms seems to be a standard part value.


The current tap is where I start to do some guessing. Looking at the tube datasheet for a 6V6 - Single Ended Class A, it says the max current draw is 45mA. 80% of that is 36mA so that is the value I used. I used 5mA for the second current tap for the 12AX7.


When I simulate the voltages I get 380V RMS at C2 and 270V RMS at C3. According to the schematic, those values should be 305V and 260V respectively.


So while there are no errors in this case, the voltages seem off so perhaps I'm not entering the proper values.

[dbishopbliss]

For the Supro 1690T, the schematic shows a 20uF, 10K, 10uF, 10K, 10uF filtering. I am using 88mA for the current based on two 6L6 tubes for I1. Then I am using 15mA for the current of the three 12AX7 tubes in the preamp, tremolo, phase inverter. I measured the actual values on my transformer.


When I run the simulation I get a dialog that says, "Warning. A current sink has pulled the voltage below zero for more than 5 mains cycles, at time 0.08..."


The voltages calculated are in the millivolts so I think something is obviously wrong.


https://el34world.com/charts/Schematics/files/Supro/Supro_1690t.pdf

[PRR]

> ..a simple model of the 5E3 Champ.

Well, much to chew. The 5E3 seems to be a DeLuxe, maybe you typed 5E1?

A quick search brought up Kuehnel's analysis of 5E1, which assumes 250VAC not 325V. And in good 1953 style, draws right on the 5Y3 curves. (Which have a "typical" value of PT resistance assumed-in.)

I don't know which of these numbers is right. But you don't care.

> A current sink has pulled the voltage...

This "current sink" is an unrealistic part. An ideal current sink/source will pull the voltage negative if it wants; no common part does, and tubes don't come close. In fact tubes are much more like resistors. If we pencil 19V/470= 40.4mA, and 290V, we have 7,200 Ohms.

The 12AX7s can't be 5mA per triode, since they have 100k resistors in series, and you don't have 100k*5mA= 500 Volts. Probably not 5mA per bottle. Kuehnel figures more like 1mA per triode.

The Supro runs 12AX7 with 270k resistors so certainly sucks less current per, surely under 5mA for all 6 triodes. And 6L6 at 88mA idle and say 350V is like a 3.5k to 4k load. Use the 5Y3 suggested PT resistances and 310VAC, we get a sweet 360V (plus). Tack on the screens/preamp string of the Supro, some wild guesses and plain Ohms Law gets numbers which should be more than close enough.

[dbishopbliss]

The 5E3 seems to be a DeLuxe, maybe you typed 5E1?

I meant 5E1... corrected the original post.

A quick search brought up Kuehnel's analysis of 5E1, which assumes 250VAC not 325V. And in good 1953 style, draws right on the 5Y3 curves. (Which have a "typical" value of PT resistance assumed-in.)

I don't know which of these numbers is right. But you don't care.

I did a search for "5E1 power transformer" and got the Edcor 5E1 site that specifies the PT Secondary 1 as 650V (325-0-325). I checked the Hammond site this morning and they recommend the 290AX which has dual Secondary 1 650/550Vct @ 100 mA. Mojo has a 5F1 PT secondary of 580V (290-0290). I guess I need to look around a bit more before plugging in numbers.

This "current sink" is an unrealistic part. An ideal current sink/source will pull the voltage negative if it wants; no common part does, and tubes don't come close. In fact tubes are much more like resistors.

Many years ago someone posted an article that said using a current sink was the way to go so that is what I have always done. It seemed to get me close to the numbers I wanted, but I think I only used it for solid state bridge rectifiers. Maybe that is why it worked. 

If we pencil 19V/470= 40.4mA, and 290V, we have 7,200 Ohms.

This is the good stuff... You know this stuff so well you don't even think about it. I'm going to write it out long-hand so I remember where these numbers came from. I see 19V is the cathode voltage written on the schematic, 470 is the cathode resisitor. Ohm's law: I = V / R... that gives us the current. At least a ballpark current. Now look at the curves for the 5Y3 rectifier. Draw a line a the 40.4mA and see where it crosses the 250VAC curve and then look to the left to get the output voltage. Kuehnel does a lot more work and comes up with 42mA load but the end result is still about 290V. Use Ohm's law again: 290V / 40.4mA = 7200 Ohms.


All that said... when I plug in the numbers to PSUD the voltage at C1 is pretty much the expected value (319 vs 320) but the voltage at C3 is 78V versus the expected value of 260V.

Is PSUD only accurate for estimating the first voltage stage or are there additional values I should be plugging in?

[sluckey]

You have the R2 (7200Ω) load in the wrong place and that's why C3 voltage is so low. That 7200Ω is the load of the power tube and should be across C2, not C3. The load resistance at C3 is much, much higher, like hundreds of KΩ.

[dbishopbliss]

You have the R2 (7200Ω) load in the wrong place and that's why C3 voltage is so low. That 7200Ω is the load of the power tube and should be across C2, not C3. The load resistance at C3 is much, much higher, like hundreds of KΩ.

I don't see a way to enter multiple loads at different points the way you can enter multiple current sinks. Perhaps I need to create separate models for the poweramp and preamp sections. If I enter a load of 78K I get the expected voltage at C3 - I just played with the values until I landed on the right voltage. Now to see if I can figure out what the value "should" be based on the design.

[dbishopbliss]

I went back with some revised values using the current sink method of modeling the power supply. I used 275-0-275V of the Hammond PT, the 40.4mA based on the cathode voltage/cathode resistor and 1.2mA for each triode in the 12AX7 at 250V from the datasheet.


C1: 318 simulated vs 320 schematic
C2: 311 simulated vs 305 schematic
C3: 258 simulated vs 260 schematic


Is this just coincidence or have I figured out the proper values to use for my model?

[shooter]

current sink method

a "sink" is = load = tube current draw, or transistor current draw or........

[sluckey]

Don't know what you have figured out but your circuit now has some believable current distribution. 

I still don't know what value this exercise has for your Supro project. You trust the Supro schematic for 99% of the circuitry. Why don't you trust the schematic values for the B+ dropping resistors?

And what about the Supro project? Has that stalled? I thought it was an interesting project until you got a wheel in the ditch.

[dbishopbliss]

And what about the Supro project? Has that stalled? I thought it was an interesting project until you got a wheel in the ditch.

I ordered some faceplates and backplates for the chassis. I need to put those on before I do the final wiring otherwise I will have to unsolder/resolder a bunch of components to put them on. They should be here tomorrow so I'm hoping to get things finished up this weekend or next weekend at the latest.

I still don't know what value this exercise has for your Supro project. You trust the Supro schematic for 99% of the circuitry. Why don't you trust the schematic values for the B+ dropping resistors?

While I was waiting I started going down the rabbit hole because the schematic says use a 10 Watt resistor in the first RC filter. I was hoping to string the resistor between the legs of the can capacitor but the size of the large resistor made it awkward. It seems like most amps use 3 Watt resistors so I was hoping to determine if the schematic was using a way over-spec'd part or if a 10W resistor was necessary.  When I started getting warnings and simulations of voltages in the single digits it made me doubt the schematic. Now that I know I can ignore the warning and I have entered numbers that are more correct I am seeing voltages that seem more reasonable.



Based on all that it seems the voltage drop across R1 will be between 80V and 100V depending on the actual current draw. Based on that it would seem 100*100/10000 = 1W so a 3W resistor should work just fine and will make things easier to assemble.


Does that make sense?

[dbishopbliss]

By the way... here is the latest model using the actual measured voltages for my transformer and the capacitor values I will actually be using.

[sluckey]

I ordered some faceplates and backplates for the chassis.

Duh! Now I remember.   

10K/3W will probably be OK, but since there's a bit of uncertainty, I would make it a high priority to measure the actual voltage dropped across that resistor soon after powering up. Calculate the power and if it's no more than 1.5 watts, I'd let it ride. If more than 1.5 watts, replace with a resistor that is at least twice the actual wattage.
 

[PRR]

....I don't see a way to enter multiple loads at different points ...

Yes, that is a shortcoming of PSUD. So? Factoring in AC-to-DC conversion, transformer loss, rectifier loss, ripple... is the hard part. Once you have semi-smooth DC, you CAN figure the small-stuff on a napkin. Or just trust the original designer.