Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: nexy_sm on October 27, 2017, 02:42:11 am
-
Hi all guys,
this is my first post on the forum. I have found a tube amp from Diastone. It is in a solid condition so I decided to repair it. Unfortunatelly I couldnt find any single photo or schematic of the Diastone amps (who knows what that really was ....). I opened it and started wth redrawing of the schematic. At some point I realized that it is a clone of the Fender '65 Deluxe Reverb, although they use EL34 in the output PP stage. I have not so much experience in repairing tube amps, I have just studied electrical engineering an dI am interested. So, for biasing G1 of the EL34 they use negative voltage, but the circuit is somehow very strange and I think it doesnt have any purpose. Here is the schematic:
(http://Screenshot_2017-10-27_09-07-17.png)
In my opinion, potentionmeter could only change potential of one gate. Second, there are now capacitors to filter out audio signal from the negative power supply. It might be thta I am wrong but I would gladly hear another opinioin.
-
i A$$sume that they wanted you to be able to balance the tubes current flow. one valve is "fixed" at whatever B- the supply delivers, the other valve can be "tuned" to match the current flow of the "fixed" valve. with such an arrangement, one assumes that the "fixed" valve will bias at or near the target B- proportional to Vg2 for the desired bias current. if you happen to buy a hot pair of valves, then you don't have any options to bias both valves colder if that is needed.
SIM of that circuit with the POT @ 50% indicated that the "tuned" valve B- is dead even with the "fixed" B-.
--pete
-
In my opinion, potentionmeter could only change potential of one gate. Second, there are now capacitors to filter out audio signal from the negative power supply.
You are exactly right. It's called a bias balance pot or sometimes called tube matching. The purpose is to set the bias point of the adjustable tube to be the same as the bias point of the non-adjustable tube. Most folks don't like this circuit and will modify it to provide adjustable bias to both tubes using the older AB763 circuit.
Your circuit is exactly the same as this Super Reverb AB568 amp...
http://el34world.com/charts/Schematics/files/fender/Fender_super_reverb_ab568_schem.pdf (http://el34world.com/charts/Schematics/files/fender/Fender_super_reverb_ab568_schem.pdf)
And here's the older, preferred AB763 circuit...
http://el34world.com/charts/Schematics/files/fender/Fender_super_reverb_ab763_schem.pdf (http://el34world.com/charts/Schematics/files/fender/Fender_super_reverb_ab763_schem.pdf)
-
Thank you for your answers. I think I will go in direction of modifying this circuit.
SIM of that circuit with the POT @ 50% indicated that the "tuned" valve B- is dead even with the "fixed" B-.
--pete
Could you please explain a bit? Really dead?
Could you please tell me what was the procedure to tune such PP stage?
-
"Dead even" means "exactly the same".
The procedure would be to measure the plate current of the non-adjustable tube. Then monitor the plate current of the adjustable tube and adjust the bias pot to give the same current reading as measured for the non-adjustable tube.
Refer to the schematic for the AB568 circuit... Set meter to measure DC voltage. Put one meter probe on pin 8 (cathode) of one output tube and put the other meter probe on pin 8 (cathode) of the other output tube. Adjust the bias pot for exactly zero volts. Both tubes will be conducting exactly the same (dead even) when the meter reads zero volts.
-
So, by looking this schematics it is probably wrong what I wrote in the topic's name. All schematics are principally the same ...
-
None of the various Deluxe Reverb models ever used that bias circuit. But the AB568 and AC568 models did. These include Bandmaster, Bassman, Super Reverb, Vibrolux Reverb, Twin Reverb, and Showman. You can see all these schematics by following the link to the Schematics Library found at the bottom of this page.
-
I see. Thank you for your help. Do you recommend changing this bias circuit?
-
Most folks don't like this circuit and will modify it to provide adjustable bias to both tubes using the older AB763 circuit.
-
Welcome to the forum nexy sm :icon_biggrin:
-
> I have just studied electrical engineering
Networks 101. You should be able to take any mess of resistors and derive all voltages/currents.
Because I am a dim-brain, I like to draw the circuit(s) as a Voltage Divider(s), top-bottom to keep my mind straight.
The "gate" (Grid) takes "zero" current and may be neglected. (But meters suck non-zero current which may have to be accounted to reconcile meter readings.)
The straightened-out voltage divider shows that the mid-turn puts the same voltage on both grids, and turning the pot puts different voltage on one tube. For hypothetical -100V into the dividers, from -100V to -69.2V, while the other hangs a constant -84.6V.
This is OK if you have crates of tubes of very consistent bias (which is not critical) but want to trim-out the balance (possibly to trim the idle buzz near none).
With today's erratic tubes we often want to trim bias on both/all tubes. Such plans can be found everywhere and I will not sketch it for you.
> "fixed" B-
Some old grumps would say A=filament, B=plate, C=grid, the three batteries in old-old radios. Thus we are talking about C feed. I agree that few of the younger crowd know what we mutter about.
-
Hi,
PRR thank you for your answer. I was reading a little bit and decided to try to follow the datasheet of the EL34 and particularly the operating conditions for push-pull setup, described on the page D2.
Here I would like to ask some fundamental questions.
1. As far as I have understood, for guitar applications, it should be enough to match the quiescent current through the output valves. I am planning to do this alignment by independently changing negative voltage on the gates.
2. The output transformer on my amp has 4 Ohm secondary while on the primary side it is written 4073. Does it make any sense to be 4kOhm?
Thank you
-
...The output transformer on my amp has 4 Ohm secondary while on the primary side it is written 4073. Does it make any sense to be 4kOhm?
That OT has a Pr:Sec impedance ratio of 1000:1 (so a 4R speaker on the secondary will reflect 4k plate-to-plate on the primary)
Your typical EL34 data sheet shows that the EL34 maximum rated plate dissipation is 25W. Running a 25W plate at say 420V on the plate in centre-bias Class A calls for a reflected load of ~7k. Running a pair of EL34 in push-pull halves this to ~3k5 for the Class A part of the signal cycle, which will kick down to 1k7 when the B load line kicks in (in Class AB1).
So a 4k plate-to-plate reflected load leaves you in safe territory for a pair of EL34s in push-pull Class AB1
-
It might be that I should not follow the datasheet, cause my secondary HT voltage is 516 V. The datasheet data I wanted to follow recommends HT 400V. Eating those 116V would be probably just the waste of energy. So, my plan is to move Vak somewhere between 475V and 500V. I will add cathode resistor (together with a big capacitor) appropriate for the Vak and anode current of about 30mA. This will also require moving Vg1 somewhere between -40V and -50V, or even beyond -50, but that I can assume from the datasheet transfer characteristic lines.
-
4k plate to plate load will be manageable for a plate voltage of 500ish. You might want to keep the screens at 380 to 420 if you can, and/or bias the tubes closer to Class B than Class A. (e.g. In Marshall Plexi's with Va around 500 and Vg2 around 493, I prefer to get the tube current around 25mA). Make sure you have a bias voltage range between -45V and -55V if using fixed bias. If cathode biasing, it will tend to pull the plate voltage down a bit.
The type of tubes you use matters at 500V. If you can find some NOS Svetlana Winged Cs (not the New Sensor rebranded ones), they will be fine at those voltages. I've had some luck with NOS Sovtek EL34G (which are still relatively cheap) where I manage to find some pairs of matching lowish current ones, although the quality of these is typically erratic (which is why I order a dozen at a time and match them myself).
If using modern production EL34s, be careful about which ones you select. They don't all hold up to 500V on the plates that well. ST EL34B seem to be okay. Lord Valve (an EL34 guru) recommends treating modern production EL34s as having a nominal 20W plate dissipation (rather than the 25W stated on old datasheets). I tend to avoid the New Sensor 'reissue' range (as I've had a few turn into expensive firecrackers taking out the OTs in expensive Marshall handwired amps).
-
Here in Germany, the easiest is to buy JJ tubes online.
I have 40V for negative supply, so -56V is minimum I can go down.
-
I see that in the original design they didnt use cathode resistor.
-
Fixed biasing gives a bit more gain (all other things, i.e. supply voltage etc, being equal) because the plate-to-cathode voltage is greater. Also, it enables you to get into Class AB1, which is more efficient, and delivers more output power than what you can get with Class A push-pull. You bias the output tubes to get somewhere between about 60% and 85% of the maximum rated dissipation (with 70% being a conventional happy medium). This keeps the tubes from over dissipating under full power, and creates the bias voltage condition necessary for the B-load line to kick in when the driver signal at the control grid gets a big enough swing to create cutoff in each alternate tube.
Cathode biasing is 'automatic biasing' because if the average tube current changes, then the average voltage drop across the cathode resistor changes, and therefore the bias voltage is automatically adjusted. You can run output tubes at 100% of their rated dissipation with cathode biasing (i.e. Class A push-pull) without biasing problems (but you ideally want to double the plate-to-plate reflected load, over what you would adopt for Class AB1, to get maximum power transfer, but even then you still don't get as much power efficiency as Class AB1).