Analysis of Altec 436/438
- Thread starter DaveP
- Start date
Help Support GroupDIY Audio Forum:
Could be VK1 compensated for lack of beef in 12AT7 by using feedback
Wouldnt that lower the output rather than beef it up ?
Iv read lots of times from guys who have some good vintage limiters that the 436 is in their words junk....As I say sadly Iv never messed with one its only recently I bin reading good things, wondered if that was a beatle thing.....On the other hand Iv read great things about the VK so am intrigued to read the design seems to be floored ?
Makes me itch for my 6BC8 so I can try it.....Im hoping its good so I can have another flavour limiter.
Wouldnt that lower the output rather than beef it up ?
Iv read lots of times from guys who have some good vintage limiters that the 436 is in their words junk....As I say sadly Iv never messed with one its only recently I bin reading good things, wondered if that was a beatle thing.....On the other hand Iv read great things about the VK so am intrigued to read the design seems to be floored ?
Makes me itch for my 6BC8 so I can try it.....Im hoping its good so I can have another flavour limiter.
skal1
Well-known member
hey gray 0
give me a ring apples and pears
skal1
give me a ring apples and pears
skal1
Hi skal1,
It will work but the tube won't have the same operating point. In the original circuit the cathode resistor is 34 ohms and shared, so just to separate them and leave the cathodes at the same voltage would need 68 ohms for each tube (half current x double resistance = same voltage) You have a 200 ohm pot plus the parallel resistance of the meter and its zero pot all in series with the cathodes, so your cathode voltages are going to be higher than the original circuit. This means that the GR does not have the same range as the original circuit, that may be good or bad or not detectable but its best to know what your mods are doing I guess. Incidently, I have found in my circuit testing that the gain is higher when the resistor is shared than when it has the equivalent single resistors, This is because the ac cancels in the resistor and the effect is like having a by-pass cap.
best
Dave P
It will work but the tube won't have the same operating point. In the original circuit the cathode resistor is 34 ohms and shared, so just to separate them and leave the cathodes at the same voltage would need 68 ohms for each tube (half current x double resistance = same voltage) You have a 200 ohm pot plus the parallel resistance of the meter and its zero pot all in series with the cathodes, so your cathode voltages are going to be higher than the original circuit. This means that the GR does not have the same range as the original circuit, that may be good or bad or not detectable but its best to know what your mods are doing I guess. Incidently, I have found in my circuit testing that the gain is higher when the resistor is shared than when it has the equivalent single resistors, This is because the ac cancels in the resistor and the effect is like having a by-pass cap.
best
Dave P
skal1
Well-known member
Hey
thanks Dave p
skal1
thanks Dave p
skal1
KrIVIUM2323
Well-known member
Hi Miha.
Ok but... With input attenuator at -30db an output attenuator to 0db you should be at unity gain. So you can suppose and guess gain of a final stage. As in the Altec where when you pull out 6al5 you have a nice preamp. For sidechain i'm ok with you it cannot give you a clue of how it interact (depending mainly of plate resistance// impedance of sidechain) as for GR tube where biasing is an important point too. When you remove R102/103 you stop negative feedback, so you give 5db more amplification so you've got more output to drive sidechain= more GR.
For me R100,101 are grid resistor, and as far as i understand for single ended (cathode grounded) biasing is given by cathode resistor not by grid resistor. Adding a negative DC to grid (and to signal ac input) vary grid biasing so varying GM in the GR tube.
It's not totally true... They're some example of class A/B design in preamps (first to mind 1081 Neve, but i think many more example could be found and maybe in comp but as you i never encountered one designed this way!
...). I agree with you about efficiency but it's a choice of design, and many more parameters are involved in design choices: price, simplicity, etc, etc... We could imagine a push pull design in classB where GR tube work on one half of signal and the same with the output tube. We could have twice more 'power' from output tube (at lesser price) making drive of sidechain easier...At cost of crossover distortion and need of perfect pairing of tubes...
So this number can't simply tell you if tube is up to the task for, say, 20-25dB GR. It depends on output tube, SC and GR tube, it's all related and pretty complicated
Ok but... With input attenuator at -30db an output attenuator to 0db you should be at unity gain. So you can suppose and guess gain of a final stage. As in the Altec where when you pull out 6al5 you have a nice preamp. For sidechain i'm ok with you it cannot give you a clue of how it interact (depending mainly of plate resistance// impedance of sidechain) as for GR tube where biasing is an important point too. When you remove R102/103 you stop negative feedback, so you give 5db more amplification so you've got more output to drive sidechain= more GR.
For me R100,101 are grid resistor, and as far as i understand for single ended (cathode grounded) biasing is given by cathode resistor not by grid resistor. Adding a negative DC to grid (and to signal ac input) vary grid biasing so varying GM in the GR tube.
It's not totally true... They're some example of class A/B design in preamps (first to mind 1081 Neve, but i think many more example could be found and maybe in comp but as you i never encountered one designed this way!
...). I agree with you about efficiency but it's a choice of design, and many more parameters are involved in design choices: price, simplicity, etc, etc... We could imagine a push pull design in classB where GR tube work on one half of signal and the same with the output tube. We could have twice more 'power' from output tube (at lesser price) making drive of sidechain easier...At cost of crossover distortion and need of perfect pairing of tubes... 
DaveP said:
Dave,
How this variation of -10volts in the 6BC8 chage its gain???....and How I can get a better balance in the halves of the tubes???.....in the real life the tubes halves never will have the same spec.!!......Someone named a balance circuit mod in another post before in order to balance the halves of tube.
Thanks a lot,
Opacheco.
I was going to use the cathode version, but tracking down that 250 pot proved to be really hard, so in respect to availability, the plate version is loads more convenient. As for which is the better solution circuit wise, I'm clueless.
> The 436 DC feedback part is that theR102 R103 100K resistors?
Not DC feedback. These come from the output transformer, which is nearly zero-impedance at DC.
These are audio feedback. They trade 12AT7 gain for lower distortion and better consistency. To a rough approximation: 12AT7 with heavy load gives gain near 35, so the 100K resistors "look like" 3K to the input tube's plates. This is a significant change from the obvious 22K loads. How this plays out, I do not know.
> if the GR stage is working class A or B.
Well, who cares?
For small signals it is surely A. Yes, if cathodes are grounded there is zero bias, but with no GR the input level MUST be quite small, under 100mV. A low-Z source can drive 100mV into a zero-bias grid. In practice most designs have just enough cathode resistance (or other idle bias) to have some slight grid-cathode bias for small signals.
For high input levels a limiter will apply negative grid bias. Vary-gain tube current is reduced. As long as the static current exceeds the peak signal current, we are still in class A.
Overload happens when the GR bias is pushed so far that static current falls lower than signal current. Interestingly, distortion is not gross even when one side cuts-off a little, as long as cathode resistance is small. So we can push somewhat into AB. Distortion gets awful (and peak output may rise) as we go through the theoretical class B point into class C.
Not DC feedback. These come from the output transformer, which is nearly zero-impedance at DC.
These are audio feedback. They trade 12AT7 gain for lower distortion and better consistency. To a rough approximation: 12AT7 with heavy load gives gain near 35, so the 100K resistors "look like" 3K to the input tube's plates. This is a significant change from the obvious 22K loads. How this plays out, I do not know.
> if the GR stage is working class A or B.
Well, who cares?
For small signals it is surely A. Yes, if cathodes are grounded there is zero bias, but with no GR the input level MUST be quite small, under 100mV. A low-Z source can drive 100mV into a zero-bias grid. In practice most designs have just enough cathode resistance (or other idle bias) to have some slight grid-cathode bias for small signals.
For high input levels a limiter will apply negative grid bias. Vary-gain tube current is reduced. As long as the static current exceeds the peak signal current, we are still in class A.
Overload happens when the GR bias is pushed so far that static current falls lower than signal current. Interestingly, distortion is not gross even when one side cuts-off a little, as long as cathode resistance is small. So we can push somewhat into AB. Distortion gets awful (and peak output may rise) as we go through the theoretical class B point into class C.
eskimo said:
Ey man!!....that's have a great explanation too!!!.......Attack time and Relese Time Network is so clear. This unit is very similar to the 436C.
But which one mod could we do in order to upgrade this great unit??....What EMI did inside of this Unit(all people talk about this but anybody knows the EMI TOP SECRET MOD)???
Thanks a lot,
Opacheco.
A few people here have been around the block with this one. My 438c clone (with mic preamp front end) has been in daily use in our studio for about 4 years now and has received very favourable comments - it works particularly well with dynamic mics for vocals, and on more extreme settings for trashing snare drums and such nonsense. It does have a particular character, and is not suitable for everything and everyone.
I went through several cycles of tweaking but ended up back at the original design, which sounded best to my ears. I bought a few tubes and selected for least thump.
If you're scratch-building a clone, my advice would be to build from high quality components and follow the original schematic, and take the time to measure and match up resistors and caps as closely as you can for the two sides. If you want to mod it after then at least you have a benchmark. Also, I used AC heating throughout, and it is dead quiet
Those 'emi'-ish mods at the input stage look like they are selected for the original transformer. If you're using a modern replacement then you may need different values or nothing at all!
I went through several cycles of tweaking but ended up back at the original design, which sounded best to my ears. I bought a few tubes and selected for least thump.
If you're scratch-building a clone, my advice would be to build from high quality components and follow the original schematic, and take the time to measure and match up resistors and caps as closely as you can for the two sides. If you want to mod it after then at least you have a benchmark. Also, I used AC heating throughout, and it is dead quiet
Those 'emi'-ish mods at the input stage look like they are selected for the original transformer. If you're using a modern replacement then you may need different values or nothing at all!
I have a PDF somewhere that lists the EMI mods....I like all of us have searched and found so called EMI mods but all the ones I found so far are disproved tho handy mods perhaps but this PDF is interesting not all details and component values there sadly but interesting some of the mods I dont understand & have posted related questions but had no responses....I will try to find it if not il find where i posted the questions I needed help with...
MagnetoSound
Well-known member
gary o said:
Discredited, not disproved.
