Stereo Amp build based on Hammond organ circuit

Is this what it looks like?
If capacitors are identical no resistors would be needed but since they have +/- 20% tolerance they will charge to different voltages based on their capacitance differences plus aging. Did I understand it correctly?
I'm not sure how to calculate resistor's power rating though


capacitors-in-series-with-resistor-voltage-divider.svg
 
Yep, thats how it would go.

Its not so much about the value as the internal leakage.

Pdiss on the resistors would be B+ squared / 470k / 2. Probably you want 1 watt parts there. If you were dealing with 600v, each resistor would dissipate 0.38 watts, so a 1w part would be a reasonable de-rating.
 
I was first thinking to shrink pairs of capacitors reversed to each other but I'm sure those resistors will be getting warm and need to be vented. Correct?

Another question. Capacitor ripple current. Higher better or not going to be smoothing well if too high? I'm thinking to use EEU-EE2V820S Panasonic caps.

Thanks for your help!
 
If they're dissipating under a watt of heat it won't get that warm, but I would not want them pressed right against the body of the cap. No point in pumping more heat in than absolutely needed.

Higher ripple current is mostly what the cap will survive. They won't filter any better, but if you have an overly low rating the cap won't live long.
 
So, I suppose 640mA @120Hz would be better then 540mA @ 120Hz, correct? Or it is an overkill?
 
I generally like as much as I can get and still have it fit. This is one of those cases where you can't over-rate it too much.
 
Hi, so my comments to this thread:

1. If You want to use this construction for a museum - leave it as it is. Looks so nice in the dark when the tubes are shining.
2. If You want to use this construction as a dailiydriver for listening, i prefer to take this rectifier tubes out and replace them with silicone diodes briged with 0,01uF Caps.
3. Double the value of main - filtercaps.

Back in the fifties the constructors had no other choice than using rectifier-tubes, if You bring them silicone diodes, they will make You vice president of their companies and bring You her daughters for marry !

The main voltage after the diodes may be connected directly to center of the outputtransformers without any resistors or anything - this is where your outputpower comes from.
Ripplevoltage is no problem at this point because the outputtransformers in PP-amps works as ckokes - thats why they are getting warm after some time of playing, not from little bias current.
After the B+ supply of endstage you can work with rc or lc assemblies for the preamp supply, but it is an easy job because they dont need much current. Bring the pre -tubes to DC-filament voltage is also a good way, did this with my pioneer sx- 800 tubereceiver, zero-hum at the speakers was the result. ( but with 10Volts ripple at the main B+ Voltage! )

Good luck
Juergen
 
I can't say anything bad about tube rectifiers. Seems like a lot of amps used them with no problem.The guitar guys love them. They have limits but used correctl your OK.
 
sp58 said:
Hi, so my comments to this thread:

1. If You want to use this construction for a museum - leave it as it is. Looks so nice in the dark when the tubes are shining.
2. If You want to use this construction as a dailiydriver for listening, i prefer to take this rectifier tubes out and replace them with silicone diodes briged with 0,01uF Caps.
3. Double the value of main - filtercaps.

Back in the fifties the constructors had no other choice than using rectifier-tubes, if You bring them silicone diodes, they will make You vice president of their companies and bring You her daughters for marry !

The main voltage after the diodes may be connected directly to center of the outputtransformers without any resistors or anything - this is where your outputpower comes from.
Ripplevoltage is no problem at this point because the outputtransformers in PP-amps works as ckokes - thats why they are getting warm after some time of playing, not from little bias current.
After the B+ supply of endstage you can work with rc or lc assemblies for the preamp supply, but it is an easy job because they dont need much current. Bring the pre -tubes to DC-filament voltage is also a good way, did this with my pioneer sx- 800 tubereceiver, zero-hum at the speakers was the result. ( but with 10Volts ripple at the main B+ Voltage! )

Good luck
Juergen
Click to expand...

Thanks for your comments.
I would probably take your advise on solid state rectification before I started this build. I'm too far into modifications and re-calculations to simply drop it. I'm determined to make this amp work correctly and I think, I'm finally on the right track. We will see.
I have learned a lot during this project and if I knew then what I know now, many of my mistakes could be avoided.

Dmitriy
 
Last edited:
This has been a great thread to follow. Lots of good trouble shooting info to learn from, save for another day.
Have you tried a 5R4 rectifier? They drop over 60 volts, rated at 250ma, tough tube.
 
Thanks for your interest. I sure have learned quite a bit. I have tried 5R4G tube. It does drop voltage to reasonable levels but, original 5AR4 is indirectly heated and will ''ramp-up'' together with the other tubes coming to life while there will be no delay with directly heated 5R4 variants and DC voltage will be available almost instantly stressing caps until tubes start conducting.
 
In the process of replacing filter caps of higher voltage rating I decided to remove inactive (except for filaments) 6CA4 rectifier tube and place 18uf/630v Solen film cap into the can and mount it on the top. This poses a potential problem. My filament voltage with all 7 tubes connected is 6.39vac. Removing rectifier will bump it up-to about 6.6vac or so. Am I safe with this voltage on filaments? 2 x CL-80 already in place. One on Line and another on Neutral.
Another thought. @sp58 suggested converting driver tube filaments to DC. I'm sure that will help with taming hum even further.

Questions:
1. Can this be done T-ing off output tube filaments and convert output to DC?
2. Bridge rectifier with RC input?
3. My filaments are raised with 25vdc now. Will that need to be removed and filaments balanced to the ground through trimmer?
4. If filaments are referenced to ground I could use FW rectifier option, correct?
5. Will rectifier and 2 driver tube heaters (0.6A) pull AC down enough to have 6.3vac on Output tubes?

I tried to model this in PSUD2 with no luck. I can't seem to get DC clean enough at 6vdc output. Any suggestions?
Thanks
 
grindfix said:
In the process of replacing filter caps of higher voltage rating I decided to remove inactive (except for filaments) 6CA4 rectifier tube and place 18uf/630v Solen film cap into the can and mount it on the top. This poses a potential problem. My filament voltage with all 7 tubes connected is 6.39vac. Removing rectifier will bump it up-to about 6.6vac or so. Am I safe with this voltage on filaments? 2 x CL-80 already in place. One on Line and another on Neutral.
Another thought. @sp58 suggested converting driver tube filaments to DC. I'm sure that will help with taming hum even further.

Questions:
1. Can this be done T-ing off output tube filaments and convert output to DC?
2. Bridge rectifier with RC input?
3. My filaments are raised with 25vdc now. Will that need to be removed and filaments balanced to the ground through trimmer?
4. If filaments are referenced to ground I could use FW rectifier option, correct?
5. Will rectifier and 2 driver tube heaters (0.6A) pull AC down enough to have 6.3vac on Output tubes?

I tried to model this in PSUD2 with no luck. I can't seem to get DC clean enough at 6vdc output. Any suggestions?
Thanks
Click to expand...
Frankly, using DCV for the input and driver tubes is usually unnecessary, but in some high gain circuits it maybe necessary.
You were worried about having 6.6 vac filaments but that is nothing to be concerned about and within 5% specs.
You can tap off the heater circuit and rectify for DCV this will use more current than tube filaments due to losses. It would be around 40% more current than the AC use. So if filament is around 1.0A then after rectification for DC and powering the tubes used would be 1.4A maybe less. Remember you only need 1 ground in the heater circuit and more than 1 would create a short.
As i explained if you rectify a part of the fil circuit that would incur an increase in overall filament current thus reducing the overall voltage. If you full wave rectify your DCV will be about 0.8 x vac.
It is a straightforward procedure, take the heater circuit full wave rectify add a cap of around 1000uf x 25vdc and there you go. I have seen circuits with more and much less capacitance, so , it doesn't seem very critical.
 
I finally had enough time to finish installing new caps, rearrange grounds and adjust EFB board resistors to suit my application. Huge thanks to Dave G for his patient explanation and guidance.

I took full load test and adjusted AC balance for even clipping. Thank you Gadget. I replace 2.7k resistors with 5k trimmers and it was enough range to get it perfect.

This thing puts out nearly 30wpc with both channels driven!
Amp is much much quieter. Now tested with Cornwalls and can say that amplifier vibration through the table is louder then slight hum through speakers. Amp is upside down resting on its power transformer for now. My filaments are still AC powered. I might still rectify filaments for driver tubes later.
One thing I'm noticing is that mids seem a bit forward and little harsh. Bass is not as deep as well. Highs are nicely detailed but not abusive.
I have Russian K40 coupling caps in it. Could it be that they, as well as new tubes just need some time to break-in?
How much time is typically needed? I could play into dummy loads for a while to do that, right?

Can frequency tests be done without Db meter? Can someone point to down to earth tutorial materials for amplifier performance testing? I have scope, RMS meter and signal generator. Possible?
 
you need a voltmeter for frequency testing, but it has to be one that is accurate across the audio range. Not all meters are. You'll also need a generator that will give the same output level across the range and produce a reasonable sine output. Phones don't make good sig gens, they get pretty nasty on the output at higher frequencies.

The volts you can turn into DB with math, or if you PM me I can email you a spreadsheet that will do it for you. Its just an Excel sheet that takes volts and turns it into db with some math functions and it will even spit out a frequency response graph based on it. The nuts and bolts of it are you need a reference point. Typically I set the output for 1 watt at 1 khz and record the voltage. Thats 0db. From there I do a range of frequencies between 20 and 20khz and record the voltages. The spreadsheet does the rest.
 
gadget73 said:
you need a voltmeter for frequency testing, but it has to be one that is accurate across the audio range. Not all meters are. You'll also need a generator that will give the same output level across the range and produce a reasonable sine output. Phones don't make good sig gens, they get pretty nasty on the output at higher frequencies.

The volts you can turn into DB with math, or if you PM me I can email you a spreadsheet that will do it for you. Its just an Excel sheet that takes volts and turns it into db with some math functions and it will even spit out a frequency response graph based on it. The nuts and bolts of it are you need a reference point. Typically I set the output for 1 watt at 1 khz and record the voltage. Thats 0db. From there I do a range of frequencies between 20 and 20khz and record the voltages. The spreadsheet does the rest.
Click to expand...


Got it. Would you think the freq response could be different at different power levels? Probably more accurate at somewhere around 50%?

One thing concerns me is that driver tubes and output tube screens at idle only pull 13.8mA combined. Tested that with AMP meter in series.
Estimate using tube data suggests around 25mA. I took voltage reading across grounded resistors in cathode circuit of driver tubes and combined they pull somewhere around 5mA which leaves 8.8mA for all 4 screens.
Tubes are biased at 33mA each.
Plates 490v and screens 405v

Problem?
 
The idea of the 1w power level test is to make sure transformer limitations aren't a factor. That will give you basic circuit performance. Even a fairly lousy output transformer should give decent results if the circuit design isn't completely awful. If that isn't decent, theres no chance of higher power performance being good.

The other test you may want to do is a power bandwidth test, which is full power -1db, and you plot the response there. That will tell you how the transformers are. Its full power -1db to make sure you're far enough below the level that clipping comes in to not skew things.
 
You must log in or register to reply here.

Similar threads

Kernel86
Hammond H-112 Organ Amp
Replies
10
Views
3K
Dswankey
Dswankey
brendandoher
Picked up a Scott Copley console with the Dual Turntable
Replies
2
Views
434
Wigwam Jones
Wigwam Jones
Kennyg2209
Re-designing a Hammond H-1 Organ Amp (PSUD2 HELP)
5 6 7
Replies
125
Views
2K
trobbins
trobbins
GordonW
Small Hammond Organ for autistic child- Atlanta to Dallas TX?
Replies
2
Views
650
qpagoda
qpagoda
jasondace2b
Wurlitzer 6431 build
Replies
3
Views
573
jasondace2b
jasondace2b
Back
Top Bottom