WntrMute, would you remember what the idle current was with B+ at 406V ?
(idle current is the current your mA-meter shows when idle = no audio output, not playing anything )
(idle current is the current your mA-meter shows when idle = no audio output, not playing anything )
Ok, tnx, now just to make sure, could you please measure the exact resistances of the 150H choke and the OPT primary winding?
Not currently as the amps are in use. Maybe this weekend. I assume you meant the 6H chokes?
here is my revised schematic. The Bias adjust for the 3C24 is outlined in a red oval.
here is my revised schematic. The Bias adjust for the 3C24 is outlined in a red oval.
No, the 6H choke does not matter, you measured the B+ with 406V, and it does not matter what the choke resistance was during that measurement.
What matters is that we had 406V after the choke.
But we do not have the exact currents yet, because of unknown 3c24 grid and anode current, so i need the exact resistances of the 150H choke and OPT.
But sure, to be able to get accurate results when simulating the ps, we need to have everything as close to reality as possible.
Both 6H chokes must be measured also because theyr real resistance can depart from 150 ohm, according to datasheet, by -20/+50%..
So, measure all choke and the OPT primary resistances.
For best accuracy conditions during the resistance measurements should be as close as possible to the measuring conditions you had when measuring the voltages #43, when originally built.
Adjust the current to the same 60mA you had originally and the temperature of chokes and OPT temperature should be as close as possible to the conditions you had during #43 voltage measurement.
Make sure the 600uf caps are fully discharged before taking resistance readings!
What matters is that we had 406V after the choke.
But we do not have the exact currents yet, because of unknown 3c24 grid and anode current, so i need the exact resistances of the 150H choke and OPT.
But sure, to be able to get accurate results when simulating the ps, we need to have everything as close to reality as possible.
Both 6H chokes must be measured also because theyr real resistance can depart from 150 ohm, according to datasheet, by -20/+50%..
So, measure all choke and the OPT primary resistances.
For best accuracy conditions during the resistance measurements should be as close as possible to the measuring conditions you had when measuring the voltages #43, when originally built.
Adjust the current to the same 60mA you had originally and the temperature of chokes and OPT temperature should be as close as possible to the conditions you had during #43 voltage measurement.
Make sure the 600uf caps are fully discharged before taking resistance readings!
Last edited:
gorgon53, measuring copper winding resistance at nominal operating temperature may be too onerous for someone with little bench experience, due to the need to double check if there are any parallel circuit paths and for safety aspects.
I'd suggest that making accurate DCR measurements at ambient temp is perhaps sufficient. Then two simulation related aspects kick in. The first is that copper winding resistance increases at +4% for each 10C rise, and although sections of windings may have quite different temperatures, a rise to circa 100C may be reasonable to assess the variation that could be experienced. PSUD2 can then be used to assess what that variation can do to an expected result (eg. output voltage for a given load current).
There are many non-ideal part parameters that influence a PSUD2 simulation, and those variations are perhaps reasonably appreciated by varying sim part values and noting the change in result. In a full-fledged sim program, that would be done using a Monte-Carlo run where all parameters that could vary are given a tolerance and thousands of sims are done that randomly vary all those parameters. PSUD2 is very simple, so it is up to the operator to check what the influence is when say choke inductance varies between 5 and 9H (-20 to +50%), and choke resistance varies from 150R to say 200R (+30%) (there is no datasheet tolerance for DCR).
I'd suggest that making accurate DCR measurements at ambient temp is perhaps sufficient. Then two simulation related aspects kick in. The first is that copper winding resistance increases at +4% for each 10C rise, and although sections of windings may have quite different temperatures, a rise to circa 100C may be reasonable to assess the variation that could be experienced. PSUD2 can then be used to assess what that variation can do to an expected result (eg. output voltage for a given load current).
There are many non-ideal part parameters that influence a PSUD2 simulation, and those variations are perhaps reasonably appreciated by varying sim part values and noting the change in result. In a full-fledged sim program, that would be done using a Monte-Carlo run where all parameters that could vary are given a tolerance and thousands of sims are done that randomly vary all those parameters. PSUD2 is very simple, so it is up to the operator to check what the influence is when say choke inductance varies between 5 and 9H (-20 to +50%), and choke resistance varies from 150R to say 200R (+30%) (there is no datasheet tolerance for DCR).
Trobbin, tnx for your input, you are quite right that correctly measuring the resistances under actual operating conditions may be to much, considering that WntrMute may not have enough experience yet to be able to do that safely.
Probably best he measures only the cold resistances when the amp has been switched off to catch possible differences from given datasheet values.
I just recalculated with the now known (took a long time to get that info) original 60mA metered current at B+ 406V.
It turned out that the difference do to the 150H choke is quite small, because the choke current of 4.22mA gets swamped by 3C24 grid current of 19.18mA,.
The dcr of the choke with 75deg estimated temp increase would go from 3700ohm to 4785ohm, with 15.6V the chokes current would go down less
than 1mA. Total current draw would drop from 66.81 to 65.81mA
This is, given that the voltages WntrMute provided in #43 and the datasheet resistances are correct.
But something more worrying than 1mA errors is the fact that the 6S4A driver is not preheated bevor switching on the anode voltage.
This means that the 6S4A and 3C24 current draw only slowly builds up together, and during this heater delay the 600Vcap voltages could rise to much
higher levels than would be safe for them and also the EF86.
Hard to say by how much, but that could have been the reason for the Xformer fault, overvoltage on the cap led to a cap short that got suddenly cleared by self healing, currentinterruption, and the resulting voltage transient took out the transformer.
I guess at least some form of voltage clamping, a VDR, or zener, during start up is needed, what is your take on this?
Probably best he measures only the cold resistances when the amp has been switched off to catch possible differences from given datasheet values.
I just recalculated with the now known (took a long time to get that info) original 60mA metered current at B+ 406V.
It turned out that the difference do to the 150H choke is quite small, because the choke current of 4.22mA gets swamped by 3C24 grid current of 19.18mA,.
The dcr of the choke with 75deg estimated temp increase would go from 3700ohm to 4785ohm, with 15.6V the chokes current would go down less
than 1mA. Total current draw would drop from 66.81 to 65.81mA
This is, given that the voltages WntrMute provided in #43 and the datasheet resistances are correct.
But something more worrying than 1mA errors is the fact that the 6S4A driver is not preheated bevor switching on the anode voltage.
This means that the 6S4A and 3C24 current draw only slowly builds up together, and during this heater delay the 600Vcap voltages could rise to much
higher levels than would be safe for them and also the EF86.
Hard to say by how much, but that could have been the reason for the Xformer fault, overvoltage on the cap led to a cap short that got suddenly cleared by self healing, currentinterruption, and the resulting voltage transient took out the transformer.
I guess at least some form of voltage clamping, a VDR, or zener, during start up is needed, what is your take on this?
Last edited:
gorgon53, I don't normally come across this type of amp, so may miss a few things.
Looking at the Hammond 156C, that's their first datasheet I've seen for a choke that gives a clear description of how the inductance was measured - so that was pleasant. However, the spec condition is 120Vac, and with 8mAdc and 3k3 DCR, that implies 26Vdc bias. Reporting the DCR at 3.339k is a bit of a joke.
I'm not sure how much voltage swing would be anticipated across that choke. With a significantly lower Vac across the choke, the inductance could be noticeably lower.
If I try to interpret the 3C24 datasheet grid voltage curves, then for about +15V grid (although we don't know the meter voltage drop), and (60-24)=36mA plate current, I can see that the idle operating point could be ballpark Vak ~ 250-400V.
To your point about power up, the 6S4A choke may well restrain its cathode voltage down near 0V due to L/R time constant, which may delay 3C24 grid current rise a bit based on the 3C24 curves.
The 6S4A neg bias would come up quick, so the V-cap would be pinned a bit below zero, and if the 405V rail overshot 550V then we do get to the choke-input concern of initial over-shoot. 600V rated parts may just be ok - and that risk likely requires testing using a variac (starting at lower mains) to see what the rail voltage rise responses are. I've had to install a HV shunt reg in two of my PA amp restorations (https://dalmura.com.au/static/HV DC shunt load.pdf) that had choke input filtering and high HT Vac, and the circuitry had negligible bleed current at turn-on to move the choke regulation down towards the critical regulation knee. Applying any form of B+ rail voltage limiting does get complicated with issues of ss device SOA, and how any pre-load current then interacts with the choke input regulation to happily sit at a safe voltage for other parts before normal idle conditions come on.
Looking at the Hammond 156C, that's their first datasheet I've seen for a choke that gives a clear description of how the inductance was measured - so that was pleasant. However, the spec condition is 120Vac, and with 8mAdc and 3k3 DCR, that implies 26Vdc bias. Reporting the DCR at 3.339k is a bit of a joke.
I'm not sure how much voltage swing would be anticipated across that choke. With a significantly lower Vac across the choke, the inductance could be noticeably lower.
If I try to interpret the 3C24 datasheet grid voltage curves, then for about +15V grid (although we don't know the meter voltage drop), and (60-24)=36mA plate current, I can see that the idle operating point could be ballpark Vak ~ 250-400V.
To your point about power up, the 6S4A choke may well restrain its cathode voltage down near 0V due to L/R time constant, which may delay 3C24 grid current rise a bit based on the 3C24 curves.
The 6S4A neg bias would come up quick, so the V-cap would be pinned a bit below zero, and if the 405V rail overshot 550V then we do get to the choke-input concern of initial over-shoot. 600V rated parts may just be ok - and that risk likely requires testing using a variac (starting at lower mains) to see what the rail voltage rise responses are. I've had to install a HV shunt reg in two of my PA amp restorations (https://dalmura.com.au/static/HV DC shunt load.pdf) that had choke input filtering and high HT Vac, and the circuitry had negligible bleed current at turn-on to move the choke regulation down towards the critical regulation knee. Applying any form of B+ rail voltage limiting does get complicated with issues of ss device SOA, and how any pre-load current then interacts with the choke input regulation to happily sit at a safe voltage for other parts before normal idle conditions come on.
I really thank you guys with working on this so diligently! It would be difficult but not impossible to preheat the ef86 and 6S4A. I might prefer that rather than add a SS circuit I don't really understand and a solution to a problem we don't know caused the PT failure. Also remaining is the issue of getting to 405V without spending 400+ dollars and without destroying the regulation.
.
Trobbin, on the internet swirl some almost similar circuits around, i regarded them to be audiophoolery because driver cathode dumping current into the grid
of a 10k Rp transmitter tube with both electrodes connected to nothing more than fresh air, did not look to me like sound engineering
This circuit makes more sense with the choke to ground.
The datasheet i found for the Hammond 156C choke differs from yours, mine from Hammond, states 150H at 8mA, Rdc 3700 ohm,
tolerance +-15% on inductance and resistance.
If this given nominal 3700 ohm is correct (measured "cold" i assume) with +-15% this can be anything between 3145 and 4255 ohm.
We must wait until WntrMute can give us the actually measured cold resistance. Then we make some adjustments to account for resistance increase do to temperature.
According to the values given in #43 the 3C24 ground to grid voltage is 15.6V.
It follows that the dc through the cold choke is in the range of 3.67-4.22-4.96mA.
With 15.6V across, the highest dc this choke will see is 5mA.
More than than 15.6V is a no no, because the anode of the 6S4A anode is rated fora 8.5W max., and based on my calculations with the nominal values we are with 9W allready above that. (could be easely fixed, since the 3C24 dissipation of only 16W on the anode and 0.3W at the grid is not even near to Its 25W max.
I would not worry about the chokes inductance, because the tubes input impedance is so low in comparison.
Usually an A-meter voltage drop is 60 or100mV fullscale, quite unsignificant i thinck.. 3C24 anode current is close to 40mA if you deduct grid and Rfb currents.
6S4A anodecurrent is 4.22 choke + 19.18 grid = 23.4mA. Operating point 392V is what wntrMute want.
My concern was more about the 6S4A being cold heater when high voltage starts up, it will not supply sufficient current to the 3C24 grid and therefore the anode cannot limit the rising up voltage. Without loading it takes less than a second to reach 600V, the 6S4A needs 15 sec to heat.
I would have suggested a string of 5W zeners, 6 of those could surely take up 30W for a few seconds and they would drop out as soon as the 6S4A is heated up sufficiently to be able to supply current to the grid. MOV could also be an option, but those have higher tolerancess i thinck, havnt looked into it
Last edited:
WntrMute, if the 5V winding on yöur powertransformer is unused (as your circuit diagram indicates) we can use that winding to prop up your powersupply in approximatly 6V steps to the level you desire.
But before that, i would want to make sure that you installed startup voltage protection. Six 5W 91V zener diodes in series across C2 should do if some form of heatcapacity is provided (for example a block connector that can soak up heat ). Do not shorten the legs. Mount away from heatsources.
But before that, i would want to make sure that you installed startup voltage protection. Six 5W 91V zener diodes in series across C2 should do if some form of heatcapacity is provided (for example a block connector that can soak up heat ). Do not shorten the legs. Mount away from heatsources.
Last edited:
That's a great idea but I'm using that winding now for that bias adjustment circuit circled in red. I don't know if I have room for an additional transformer.
Love the out of the box thinking!
Love the out of the box thinking!
If thats so, then we need to make some changes and integrate that bias circuit into it, should not be to difficult, but not today, soon 4am here.
No need for transformers, but caps, every 6V step need a bridge rectifier and 3 or 4 caps
No need for transformers, but caps, every 6V step need a bridge rectifier and 3 or 4 caps
WntrMute, looks like your circuit diagram still needs some corrections. is there really the + side of the doubler grounded or is there a mistake in your diagram? (the doubler gives +- 6V, a total of 12V).. The 1000uF cap looks shorted, B+ looks shorted to ground, the outer connection of the 10k pot looks to be connected to B+ (should be connected across across the 1000uF i guess).
Anyway, I would have tougth there is a +- 6V regulation in respect to ground.
Please tell me the exact bias regulation range in respect to ground, is it really zero to -12V or +-6V?
Anyway, disregarding the implications of the biasing, I simmed what i had in mind and it should work even if your first choke would saturate at start up and the only load would be the 330+47k resistors.
The 5W zeners are only burdened with 5W each in case the chokes inductance plummets do to saturation and falls to 10% of the rated 6H.
Possible (because peak current is around 1A during start up) but only for a very short time. As soon as the choke L1 comes out of saturation it regaines its nominal inductance of 6H and the zener diodes heat burden will be maximal 1.2W each. And this 1.2W only during the time the tubes arent yet heated enough to take over the load. As soon as the tubes suck enough current to drop the voltage drops below ca 540V the zeners cease to conduct any current.
What I have in mind needs:
6 pcs 1N377B 5W 91V zener diodes serial connected. From ground first anode-cathode-anode-cathode a.s.o. ending with cath (the white ring) toB+.
4-6pcs rectifier diodes, any of the diodes in the 4000 serie or equivalent will do for the voltage multiplier.
4-6pcs 16V 4700uF caps, best those with low ESR, high temperatur and high ripple current capability (the types used for switched PS)
The voltage lifting is done with the caps and rectifier diodes as "Villard" voltage multiplier (look it up in wikipedia).
4 diodes + caps givs 23-24V lift
6 diodes + caps givs 32-33V lift
Voltage can be dropped with serial resistors or diodes.
smaller steps would have been possible with a fullwave bridge doubler (my original idea) but that would have ment only more diode voltage drops and 10000uF caps, so i dismissed that idea.
Anyway, I would have tougth there is a +- 6V regulation in respect to ground.
Please tell me the exact bias regulation range in respect to ground, is it really zero to -12V or +-6V?
Anyway, disregarding the implications of the biasing, I simmed what i had in mind and it should work even if your first choke would saturate at start up and the only load would be the 330+47k resistors.
The 5W zeners are only burdened with 5W each in case the chokes inductance plummets do to saturation and falls to 10% of the rated 6H.
Possible (because peak current is around 1A during start up) but only for a very short time. As soon as the choke L1 comes out of saturation it regaines its nominal inductance of 6H and the zener diodes heat burden will be maximal 1.2W each. And this 1.2W only during the time the tubes arent yet heated enough to take over the load. As soon as the tubes suck enough current to drop the voltage drops below ca 540V the zeners cease to conduct any current.
What I have in mind needs:
6 pcs 1N377B 5W 91V zener diodes serial connected. From ground first anode-cathode-anode-cathode a.s.o. ending with cath (the white ring) toB+.
4-6pcs rectifier diodes, any of the diodes in the 4000 serie or equivalent will do for the voltage multiplier.
4-6pcs 16V 4700uF caps, best those with low ESR, high temperatur and high ripple current capability (the types used for switched PS)
The voltage lifting is done with the caps and rectifier diodes as "Villard" voltage multiplier (look it up in wikipedia).
4 diodes + caps givs 23-24V lift
6 diodes + caps givs 32-33V lift
Voltage can be dropped with serial resistors or diodes.
smaller steps would have been possible with a fullwave bridge doubler (my original idea) but that would have ment only more diode voltage drops and 10000uF caps, so i dismissed that idea.
Last edited:
First question: Correct, the +side is grounded giving me up up to 12V negative applied to the grid.
Your proposition seems very interesting but I am not expert enough to picture hoe this all would work. I appreciate all the help!
Your proposition seems very interesting but I am not expert enough to picture hoe this all would work. I appreciate all the help!
I have pulled the amps out (listening to my 300Bs for a change). What other measurements /info would be helpful at this point?
Do you folks agree with the total draw of 106mA? I am looking at some chokes with 100mA current capacity depending on which way I decide to go.