Hi all,
I am trying to find a source of learning that can help me increase my understanding of valve circuits. I know basic valve theory etc but want to try and grasp a bit more 🙂
Take the circuit of my current interest - the Mullard 3.3 amp, I'm collecting parts to put one together as a bench experiment but want to try and figure out various bits.
Lets start with how do you calculate the expected voltage at the anode of the EF86 pre-amp? I know the circuit is a little odd as they have gone for a very high plate resistor of 1M but also direct-connected the output to the grid of the EL84. I see there are three stages of smoothing - R15/C9, R12/C6, R7/C3 and the output of the EZ80 is around 320v.
Maybe i am looking at it wrongly, but how would i know what wattage the 1M plate resistor should be or indeed the same for R15, R12 & R7???
Working out currents and voltage drops is easy if you have the unit on the bench but when its all on paper where do you start???
Thanks
@davek0974
Page 4 has complete build, parts list, circuit description and voltages of the 3-3.
Frank
Or the UK edition of "circuits for audio amplifiers" is a useful introduction and describes several other circuits too. Well worth getting or downloading a copy.
Sadly many of the brilliant minds have long departed, but their legacy remains. P J Walker of Quad to name but one.
To understand the black art of electronics is to understand witchcraft. Andrew.
@davek0974
Page 4 has complete build, parts list, circuit description and voltages of the 3-3.
Thanks for that, looks good.
.... And a free-to-view copy exists in the KB for guests and those that don't yet have data library access (see below). However I will point out Dave has already stated he has this book in his other thread. (see here).
CrustyTV Television Shop: Take a virtual tour
Crusty's TV/VCR Collection: View my collection
Crustys Youtube Channel: My stuff
Crusty's 70s Lounge: Take a peek
Not the easiest circuit to start with as it is that rarity in audio amps of being DC-coupled! The reason for the anode resistor being so high is that the EF86 is being run in "starvation"; the very low anode and screen currents also reduce the noise of what is already a low-noise valve. This allows the anode voltage to be at a level where the EL84 has the correct bias.
Have a look also at the treasure trove here - https://frank.pocnet.net
Not the easiest circuit to start with as it is that rarity in audio amps of being DC-coupled! The reason for the anode resistor being so high is that the EF86 is being run in "starvation"; the very low anode and screen currents also reduce the noise of what is already a low-noise valve. This allows the anode voltage to be at a level where the EL84 has the correct bias.
Have a look also at the treasure trove here - https://frank.pocnet.net
This leaves me a little confused. Mullard quote the anode of EF86 as 30V and the cathode of EL84 as 27V. Surely this is not practical? Have I missed something?
Not the easiest circuit to start with as it is that rarity in audio amps of being DC-coupled! The reason for the anode resistor being so high is that the EF86 is being run in "starvation"; the very low anode and screen currents also reduce the noise of what is already a low-noise valve. This allows the anode voltage to be at a level where the EL84 has the correct bias.
Have a look also at the treasure trove here - https://frank.pocnet.net
This leaves me a little confused. Mullard quote the anode of EF86 as 30V and the cathode of EL84 as 27V. Surely this is not practical? Have I missed something?
Yes - there's an error. Looking at http://www.r-type.org/articles/art-003h.htm the correct values are 20V on the control grid of the EL84 and 28V on the cathode of same, giving a grid bias of 8V. This means the cathode current (Ia + Ig2) of the EL84 is 50mA (the amount taken by the screen of the EF86 can be ignored).
@cathovisor Thank you for your reply. There I go, taking the maker's word for it again. Sanity (?) is restored.
Hi all,
i finally got the parts to put together my version of the Mullard 3-3 and would like some guidance on debugging it.
I only have 2nd hand valves so far as new were £££
It came up with no smoke which was good but then proceeded to blow my eardrums with massive hum. Disconnecting R8 the connection from EF86 anode to EL34 grid stops the hum and its silent, no red-plating etc.
Taking voltages i get HT of 320v, R12 is dropping 29v, R7 is dropping 49v.
There is 125v to the 1M plate load on the EF86 but its got zero on the plate - indicates it's driven hard-on?
There is also 0v on the cathode with respect to ground, 35v on G2, 1v on G1 with no input connected (seems a little odd?)
The EL84 shows 39v on the cathode, 177v on G2, 184v on the anode.
I have read through the wiring again and seems ok, the values of resistors are all ok, chances are there is a boo-boo somewhere but would like a little help in figuring out how to debug 🙂
Thanks
What happens if you short out the input ... it's input impedance is very high.
What happens if you short out the input ... it's input impedance is very high.
At the input or at the grid?
With the grid grounded the anode rises to 0.39v, G2 drops to 21.95v and the cathode shows 0.08v
Oddly, with the inter-stage connection still disconnected, touching the input creates a good hum in the output which seemed odd?? Cathode coupling via R9 maybe?
Go over your work a few times and check everything, including all resistor values.
The 1vdc on the input grid to the EF86 does indeed seem very wrong. I would expect zero volts. It would seem at first glance that your EF86 is defective. There could be internal DC leakage creating a partial short.
You should get in the region of 20vdc at the anode of the EF86 and 28vdc at the screen grid. The cathode voltage will be very low anyway, but as the anode load resistor is very high, the internal resistance of your multimeter will further reduce the reading at the anode.
The EL84 should have about 290vdc on the anode and screen grid with roughly 28vdc on the cathode. Do not operate the circuit for long without any connection to it's 1st grid via R8. You will damage the valve without proper bias, hence you have 39v on the cathode which indicates too much current through the EL84.
If all else checks ok, source another EF86. (Not a bad idea to have spares anyway, certainly handy for reference if a fault should occur).
To understand the black art of electronics is to understand witchcraft. Andrew.
Thanks, i'll delve into that area as soon as, i have three EF86's so hopefully one will be workable.
My calculation shows about 27mA through the OPT.
Sorted it - it was oscillating badly, i changed the EF86 and the EL84 and the sound changed from a hum to a machine-gun sound 🙂 Seems it was caused by having positive feedback not negative - reversing the OPT primary stopped the noise and the voltages looked a lot better - still off as i'm only running a 240-0-240 transformer, i have a new one bought especially for the 3-3 which is 300v but this is just a lash-up.
Voltages WRT ground are now:
HT = 242, after R15 = 232, after R12 = 203, after R7 = 147
EF86 anode = 27.5, g2 = 29.5, k = 0.02
EL84 anode = 215, g2 = 202, k = 29.8
It has some serious gain - un-shorting the input gives mega-hum, connecting up an iPod stops it though, i guess this is the extreme high impedance at the input? There is a small amount of hum but the build was not done with quality in mind, it is a learning test-bed at present. The OPT primary has long (12") figure-8 wire to it as its not even on the chassis 🙂 so does the speaker connections back to the tag-strips.
Can't really assess quality/tone as my junk speaker is just a 4" unit from an old midi-stack which always sounded poor 🙂 But it does play and sounds ok.
With the grid grounded the anode rises to 0.39v, G2 drops to 21.95v and the cathode shows 0.08v
The reason for that change is that the EF86 is using grid-current bias, hence the 10M resistor. I always thought this a nasty way of doing things.
I'm also a bit concerned about the bias on the EL84, as it looks rather low - only in the region of a couple of volts?
The other thing is that it shouldn't have 'loads of gain' - unless the maths of feedback fractions has changed since I was a student, at low frequency it should be in the region of 40 to 80 depending upon the value of R4 you've chosen.
@cathovisor R4 is a point - in the diagram they have a fixed OPT secondary i presume, and changing the value of th speaker means using a different R4 ??
My OPT is multi-tapped 4-8-16 ohm, what does this do to the value of R4??
When i say loads of gain, i guess i meant its extremely sensitive 😉
@cathovisor R4 is a point - in the diagram they have a fixed OPT secondary i presume, and changing the value of th speaker means using a different R4 ??
My OPT is multi-tapped 4-8-16 ohm, what does this do to the value of R4??
When i say loads of gain, i guess i meant its extremely sensitive 😉
Well, you should choose the value of resistor and transformer tap in your case to match the impedance of the loudspeaker you want to use - so if you're using an 8Ω loudspeaker, you should use the 8Ω tap on the transformer and R4 should be 100Ω.
You could, if you were feeling adventurous, use a 3-position 2-pole make-before-break rotary switch to make these changes selectable.
As for "extremely sensitive", I wouldn't call this
The comparatively high sensitivity of the amplifier (100 mV for 3 W) permits the use of all types of crystal pick-up head and allows, if required, the use of equaliser networks between the head and amplifier.
extremely sensitive?
