I assume doing this with the pre stage valves of circuit (as they are currently) won't make much difference to results?
Post #108 indicated that heater windings gave 5.9 and 6.3V with the 220V primary tap. Those heater voltage levels are likely acceptable, especially if they are confirmed to be the voltages at the EL34 heater socket terminals (perhaps when each EL34 has 6V direct across it, and not 12V for a series connection). Perhaps you still have the EL34 heaters wired in series - if so then need to confirm what the '12V' voltage is, and that each EL34 heater is actually getting 50% of 12V. If the voltage is down at say 11V then that would indicate you need to use the 240V primary tap, so that the heater voltage comes up closer to 6.3V nominal.
If you then are using the 220V primary tap then that would indicate you have a 375-0-375V secondary, which appears to then give you 356V. With that voltage, and with the post #42 schematic, I'd recommend you insert 10 ohm series resistor with each EL34 cathode, and use a measurement of dc voltage across each 10 ohm to confirm the idle current draw from each cathode. For a safe dissipation of say 20W, and a anode-cathode voltage of about 340V, each idle cathode current would be about 20W/340V = 60mA. Preferably any one EL34 does not pass more than say 65mA. The anode to cathode voltage is lower than B+ (356V) by the drop across the 130 common cathode (eg. 17V drop at 65+65mA), and a drop across the output transformer primary half-winding resistance (TBD).
If you end up using the 240V primary tap, with 415-0-415V secondary, then need to confirm the B+ voltage available under load, but need to recalculate the cathode current that keeps the dissipation safely under 25W. That is where a variac can allow operating at a lower mains voltage up until you know what each EL34 is drawing, and can confirm that you are safely under the 25W limit for each valve.
Preamp stages are unlikely to draw more than 5mA, so when compared to circa 130mA for the output stage, it is almost insignificant.
If you then are using the 220V primary tap then that would indicate you have a 375-0-375V secondary, which appears to then give you 356V. With that voltage, and with the post #42 schematic, I'd recommend you insert 10 ohm series resistor with each EL34 cathode, and use a measurement of dc voltage across each 10 ohm to confirm the idle current draw from each cathode. For a safe dissipation of say 20W, and a anode-cathode voltage of about 340V, each idle cathode current would be about 20W/340V = 60mA. Preferably any one EL34 does not pass more than say 65mA. The anode to cathode voltage is lower than B+ (356V) by the drop across the 130 common cathode (eg. 17V drop at 65+65mA), and a drop across the output transformer primary half-winding resistance (TBD).
If you end up using the 240V primary tap, with 415-0-415V secondary, then need to confirm the B+ voltage available under load, but need to recalculate the cathode current that keeps the dissipation safely under 25W. That is where a variac can allow operating at a lower mains voltage up until you know what each EL34 is drawing, and can confirm that you are safely under the 25W limit for each valve.
Preamp stages are unlikely to draw more than 5mA, so when compared to circa 130mA for the output stage, it is almost insignificant.
As said before, a Marshall'ish design were one option that I'd go if your 2 x EL34 amp were mine. Just connect the output transformer's primary CT to the first PSU filter capacitor, just as Marshall (and many others) did and does. Three ECC83's would suffice for this design. Leave the superfluous socket open.
You could also implement a Mesa Boogie'ish Dual Rectifier by adding two plain silicon 1N4007 diodes as a full wave rectifier and an additional switch that allows you to connect these diodes' common cathode to the EZ81 cathodes.
The other option were a Hiwatt design. To be honest, I'd prefer this one, as I think your PA output transformer comes more close to the HiFi'ish Partridges that Dave Reeves used than to a Marshall OT. For schematics, you might want to look at Mark Huss' site.
Best regards!
You could also implement a Mesa Boogie'ish Dual Rectifier by adding two plain silicon 1N4007 diodes as a full wave rectifier and an additional switch that allows you to connect these diodes' common cathode to the EZ81 cathodes.
The other option were a Hiwatt design. To be honest, I'd prefer this one, as I think your PA output transformer comes more close to the HiFi'ish Partridges that Dave Reeves used than to a Marshall OT. For schematics, you might want to look at Mark Huss' site.
Best regards!
Thanks for that, Kay Pirinha. Will check some hiwatt circuits out.
Just to confirm, I've checked today and the voltage selector was in the 240-250 position. So this was the winding used when I recorded 356v HT available with the pre amp nodes disconnected from HT supply
The (presumed but I'd say likely) heater output pins are giving me 5.8v-ct-5.5v AC (unloaded).
I also tested the unloaded output at the high voltage secondary (I've got the recorder tubes pulled), 383v-ct-383v
Same test repeated using the 220-230 primary winding:
Heater (presumed) 6.42v-ct-6.05v AC
High voltage 411v-ct-411v.
So these are very similar to the results mentioned from post #108 and confirms this to be correct.
Staying with the 220-230 primary then, I did the bias measurement for the EL34's.
The 2 EL34's are way apart. The plate current for one works out at 147ma! There's a 30ish voltage difference between the plate voltage of each, the one with the higher voltage has 9ohms less resistance to the OT CT when measured.
I'm guessing the anode resistor of the valve with the higher anode voltage could be adjusted to bring the plate voltages closer together?
The voltage drop measurements are miles apart, tube 2 17.08v Vs tube 2 47.15v.
Tube 1 = 23.87w plate dissipation
Tube 2 = 59.82w plate dissipation
The Cathodes have a shared 130ohm resistor (tests ok for resistance) the anodes each have their own a 100ohm resistor between themselves and the O/T HT supply.
👍
Just to confirm, I've checked today and the voltage selector was in the 240-250 position. So this was the winding used when I recorded 356v HT available with the pre amp nodes disconnected from HT supply
The (presumed but I'd say likely) heater output pins are giving me 5.8v-ct-5.5v AC (unloaded).
I also tested the unloaded output at the high voltage secondary (I've got the recorder tubes pulled), 383v-ct-383v
Same test repeated using the 220-230 primary winding:
Heater (presumed) 6.42v-ct-6.05v AC
High voltage 411v-ct-411v.
So these are very similar to the results mentioned from post #108 and confirms this to be correct.
Staying with the 220-230 primary then, I did the bias measurement for the EL34's.
The 2 EL34's are way apart. The plate current for one works out at 147ma! There's a 30ish voltage difference between the plate voltage of each, the one with the higher voltage has 9ohms less resistance to the OT CT when measured.
I'm guessing the anode resistor of the valve with the higher anode voltage could be adjusted to bring the plate voltages closer together?
The voltage drop measurements are miles apart, tube 2 17.08v Vs tube 2 47.15v.
Tube 1 = 23.87w plate dissipation
Tube 2 = 59.82w plate dissipation
The Cathodes have a shared 130ohm resistor (tests ok for resistance) the anodes each have their own a 100ohm resistor between themselves and the O/T HT supply.
I'm not using the mains transformer for heater supply, I have a separate 6v-ct-6v transformer with one side supplying the pre amp valves and the other supplying the EL34 and rectifiers.
👍
I double checked the plate voltages on the EL34's which where still the same </> a few volts. I did this because I wanted to swap the valves around. With valve 1 in valve 2's socket and vie versa, I was seeing 400 </> on both plates.
Repeated the biasing measurements again with the valves in these positions. One thing I've noticed is that the voltage of the anode, and the reading for the voltage drop jumps around quite a lot. Say the plate voltage reading for EL34 2/2 as seen bellow at 392, I seen this as low as 382. It fluctuated quite a lot up and down on all the voltage measurements for both valves, even after 'warming up'. Would this point towards rectifier valves playing up? Or is this sort of fluctuation expected?
Anyway, here's the bias measurement procedure repeated:
Way over maximum plate dissipation of 25w, assuming I've measured and calculated it properly?
Repeated the biasing measurements again with the valves in these positions. One thing I've noticed is that the voltage of the anode, and the reading for the voltage drop jumps around quite a lot. Say the plate voltage reading for EL34 2/2 as seen bellow at 392, I seen this as low as 382. It fluctuated quite a lot up and down on all the voltage measurements for both valves, even after 'warming up'. Would this point towards rectifier valves playing up? Or is this sort of fluctuation expected?
Anyway, here's the bias measurement procedure repeated:
Way over maximum plate dissipation of 25w, assuming I've measured and calculated it properly?
You need to subtract the cathode resistor voltage to calculate the real plate voltages, and you'd also subtract the screen current from the total cathode current.
Anyway, it's better to provide individual cathode resistors for each tube. Double the value of the common resistor. And if you intend to go for Marshall or Hiwatt, provide fixed bias to achieve »the« sound.
Best regards!
Anyway, it's better to provide individual cathode resistors for each tube. Double the value of the common resistor. And if you intend to go for Marshall or Hiwatt, provide fixed bias to achieve »the« sound.
Best regards!
It is often not practical to measure the anode voltage, for the purposes of calculating idle power dissipation, due sometimes to noise, and the risk of high voltages if the valve itself is noisy and causing the anode voltage to fly around. Imho, it is better to measure the output transformer DC resistance from B+ (primary CT) to each anode tap, and then calculate the nominal voltage drop across each anode winding at the measured cathode current (from the cathode sense resistor). Then subtract the cathode and anode related voltage drops for the B+ voltage to give a nominal operating anode-cathode voltage for calculating power dissipation.
The advantage of using a variac is that EL34 power dissipation can be assessed before it damages the EL34. If in fact your EL34's were dissipating upwards of 45W, and for more than a few seconds, then you may well have caused them some damage, although it is likely they were degraded already, and any path you want to take with the amp needs to cover either the purchase of replacement EL34, or a changeover to other valve types (for whatever reason, such as you have some others, or they are cheaper etc). Damage occurs within an EL34 due to the plate structure temperature rising above nominal (the red-plating affect), which then causes the metal to 'out-gas' molecules that affect the vacuum and cause various forms of damage that may or may not be recoverable with time.
There could also still be other faults in parts that are causing gross over-bias of the EL34, which is an advantage of going through what I would call an early testing phase of all parts and operation (compared to gutting the amp for starters) so that you better appreciate what gremlins exist.
If the EL34 were each dissipating 45W, then that would also grossly suppress your B+ voltage, and so misguide your comparisons.
The advantage of using a variac is that EL34 power dissipation can be assessed before it damages the EL34. If in fact your EL34's were dissipating upwards of 45W, and for more than a few seconds, then you may well have caused them some damage, although it is likely they were degraded already, and any path you want to take with the amp needs to cover either the purchase of replacement EL34, or a changeover to other valve types (for whatever reason, such as you have some others, or they are cheaper etc). Damage occurs within an EL34 due to the plate structure temperature rising above nominal (the red-plating affect), which then causes the metal to 'out-gas' molecules that affect the vacuum and cause various forms of damage that may or may not be recoverable with time.
There could also still be other faults in parts that are causing gross over-bias of the EL34, which is an advantage of going through what I would call an early testing phase of all parts and operation (compared to gutting the amp for starters) so that you better appreciate what gremlins exist.
If the EL34 were each dissipating 45W, then that would also grossly suppress your B+ voltage, and so misguide your comparisons.
Thanks, Trobbins. Would it not make sense, seen as these have already been subjected to an unsafe plate dissipation to get them within range? Any damage is already done so I'm not worried about them now. Is there any chance they may be ok by the way?
What I mean is, if it's likely they're toast, no harm keeping them in while I fault find? Would be nice if there's a possibility they might be usable in the long run but I'm willing to replace obviously 👍
I'm just giving you a heads-up about their possible condition. If you have a variac, then bring them up such that their their cathode current is not excessive, as that also avoids stressing the power transformer and output transformer from excessive current levels.
Faults that could cause excessive output stage current include leakage through the coupling caps, or bad grid leaks, First check the value of the grid-leaks (value is not shown in post #42 schematic). Checking for leakage through the 0.15uF coupling caps is best done with a variac, and removing the EL34's and measuring the dc voltage across the grid leaks as B+ is increased - but be mindful not to let any power rail voltage rise to a level to stress any filter cap voltage rating. The dc voltage across the grid leaks will rise as the B+ rises (as the coupling caps are charging through the grid leaks), but after a minute perhaps should be well below 100mV. A leaky coupling cap will bias an EL34 on and cause excessive idle current.
Faults that could cause excessive output stage current include leakage through the coupling caps, or bad grid leaks, First check the value of the grid-leaks (value is not shown in post #42 schematic). Checking for leakage through the 0.15uF coupling caps is best done with a variac, and removing the EL34's and measuring the dc voltage across the grid leaks as B+ is increased - but be mindful not to let any power rail voltage rise to a level to stress any filter cap voltage rating. The dc voltage across the grid leaks will rise as the B+ rises (as the coupling caps are charging through the grid leaks), but after a minute perhaps should be well below 100mV. A leaky coupling cap will bias an EL34 on and cause excessive idle current.
Nice one. Plenty to be going at there. I'll need to go away and digest it. Will go over what you've suggested when I'm next in front of it. Cheers.
What I find odd is the original bias test results Vs the results after switching the EL34'S socket positions. Although they're miles off spec the second time around, the second lot of results are quite different to the first.
I'm wondering about the sockets, whether poor contacts could be to blame for this? I've cleaned the chassis but I think I'll use some contact cleaner on the socket pins.
Also wondering if the primary mains transformer winding is putting out too much voltage on the secondary? I know the heater winding makes more sense at 220-230, but maybe it's putting out too much on the HT winding than the circuit was designed for? My measured mains is usually around 240.
I'm wondering about the sockets, whether poor contacts could be to blame for this? I've cleaned the chassis but I think I'll use some contact cleaner on the socket pins.
Also wondering if the primary mains transformer winding is putting out too much voltage on the secondary? I know the heater winding makes more sense at 220-230, but maybe it's putting out too much on the HT winding than the circuit was designed for? My measured mains is usually around 240.
Just mulling this over, Trobbins. The pre amp stages where not in circuit during the testing. I.e, the HT nodes where disconnected from HT supply. Wouldn't that rule out the coupling caps as there would be no DC on the pre stage valve anodes to pass through? Just a thought....
If you did not see that the EL34s were distressed, i.e. the outer plate visible through the glass glowing a cherry red, then it is possible your measurement strategy was incorrect. If you included the choke in the total resistance then some of the drop is lost in the choke itself.
Is it difficult to measure the voltage drop across the 100R resistor on each EL34 anode? You can set up the measurement, switch on the amp and switch off again to be safety conscious. This isn't the whole picture because the screen also has a contribution to dissipation, but it will get you in the ball park.
Is it difficult to measure the voltage drop across the 100R resistor on each EL34 anode? You can set up the measurement, switch on the amp and switch off again to be safety conscious. This isn't the whole picture because the screen also has a contribution to dissipation, but it will get you in the ball park.
The schematic in post #42 has an error: the 0.15uF coupling cap to 'top' EL34 should connect to pin 1 of phase inverter triode, not to the 'HT' as shown. With that PI triode removed, the 'top' 0.15uF coupling cap should develop 'HT' volts across it, as it has 212k on one side to HT, and grid leak to ground on other side - any leakage through that cap will cause some voltage drop across the grid leak resistor.
Without the PI triode then yes the 'bottom' coupling cap has no voltage across it, so can only be tested for leakage when the PI triode is in circuit and something like 100V on its cathode.
If there was no 'HT' supply voltage during your testing, then yes there should be no leakage issue to cause EL34 grid to rise (apart from gross leakage through the EL34 grid, which should be suppressed to cause no more than say 0.1V across the grid leak resistor, which for the EL34 is stated in the datasheet as 700kohm max).
Without the PI triode then yes the 'bottom' coupling cap has no voltage across it, so can only be tested for leakage when the PI triode is in circuit and something like 100V on its cathode.
If there was no 'HT' supply voltage during your testing, then yes there should be no leakage issue to cause EL34 grid to rise (apart from gross leakage through the EL34 grid, which should be suppressed to cause no more than say 0.1V across the grid leak resistor, which for the EL34 is stated in the datasheet as 700kohm max).
Thanks, TR. It's likely just a mistake on the drawing, I'll check it out. Looking at this pic (I'm working today and not in front of the amp), it looks as though it does connect directly to pin 1 of the PI. I'll check it properly and re draw an up to date sketch tonight.
I just followed a guide which give a method of calculating dissipation by measuring the impedance between the EL34 anode and pin 8 on the rectifiers. This does incorporate the choke, yes.
I can measure the drop across the anode resistor, I may do as you've suggested with the Variac? If I set it to 120v perhaps?
It has to be full B+ otherwise you are not measuring at the operating point.
Be prepared to turn off if the voltage drop is much more than expected. If the anode current should be 60mA, then the voltage drop across 100R is (from I = V / R), 100 x 0.06 = 6v.
Be prepared to turn off if the voltage drop is much more than expected. If the anode current should be 60mA, then the voltage drop across 100R is (from I = V / R), 100 x 0.06 = 6v.
Ok, I measured the V drop across each anode just now, however only the EL34'S where HT connected. I will repeat with the PI valve connected to HT?
The (probably irrelevant) voltage drop I recorded was 8.12v and 8.19v.
I re reassured the screen resistor value, this actually shows 412 ohms, not 44k as shown on the sketch. This must have been an error, I can only think it must have measured 440 something first time and I've read wrong.
The grid resistors (values not shown on the sketch) are 228k and 212k.
I will amend the sketch and repost.
The (probably irrelevant) voltage drop I recorded was 8.12v and 8.19v.
I re reassured the screen resistor value, this actually shows 412 ohms, not 44k as shown on the sketch. This must have been an error, I can only think it must have measured 440 something first time and I've read wrong.
The grid resistors (values not shown on the sketch) are 228k and 212k.
I will amend the sketch and repost.