I came across an old commercial push pull amp design which uses a TVS diode to drop 120V from a UL tab at 415V. The diode is connected between OPT UL tab and EL84 screen and is bypassed with a 0.047 uF capacitor.
The part number is Z6120. The datasheet says it is a 120V bipolar TVS diode with 1.5 KW pulse peak power. I have a few questions on this design:
1. Why use TVS diode instead of Zener diode? If TVS diode is better, why have I not seen many people use it in this specific application?
2. Can it be used to drop just 20V - 50V so that I can use TV tubes that frequently require lower screen grid? How does this approach compare against simple 1K - 2K drop resistors?
3. The Z6120 is a biploar device. The unipolar version is Z6120U. The bipolar part number may not be totally correct. Should I use bipolar or unipolar for this application?
4. What pulse peak power should I use? The original part is rated for 1.5KW. As screen grid current for EL84 is unlikely to be too high, is 1.5 KW rating overly conservative? The commonly P6KE series TVS diodes are rated for 600 watts and was suggested as a replacement for Z6120. What is your thoughts on that?
By the way, the sound of the commercial amp is highly rated. So I don't doubt that TVS diode works in the design. I want to learn more so that I can borrow the design in my own amplifier.
The part number is Z6120. The datasheet says it is a 120V bipolar TVS diode with 1.5 KW pulse peak power. I have a few questions on this design:
1. Why use TVS diode instead of Zener diode? If TVS diode is better, why have I not seen many people use it in this specific application?
2. Can it be used to drop just 20V - 50V so that I can use TV tubes that frequently require lower screen grid? How does this approach compare against simple 1K - 2K drop resistors?
3. The Z6120 is a biploar device. The unipolar version is Z6120U. The bipolar part number may not be totally correct. Should I use bipolar or unipolar for this application?
4. What pulse peak power should I use? The original part is rated for 1.5KW. As screen grid current for EL84 is unlikely to be too high, is 1.5 KW rating overly conservative? The commonly P6KE series TVS diodes are rated for 600 watts and was suggested as a replacement for Z6120. What is your thoughts on that?
By the way, the sound of the commercial amp is highly rated. So I don't doubt that TVS diode works in the design. I want to learn more so that I can borrow the design in my own amplifier.
Pulse power is irrelevant for voltage regulation using a zener or TVS.
Unipolar is all you need for voltage regulation of a DC circuit.
If its a 120V zener it can be used for regulating 120V only, not for 20V (should this need saying?)
The voltage ratings for TVS diodes are different than zeners as TVS diodes are designed not to conduct in normal use, zeners are designed to conduct in normal use.
All in all using a zener is better, its easier to get the right voltage, and the intention is clear (TVS diodes are used for protection from transient voltage spikes carrying lots of energy). But a TVS diode will work if you figure out the right voltage for the current it will be carrying (look at the curves in the datasheet).
Unipolar is all you need for voltage regulation of a DC circuit.
If its a 120V zener it can be used for regulating 120V only, not for 20V (should this need saying?)
The voltage ratings for TVS diodes are different than zeners as TVS diodes are designed not to conduct in normal use, zeners are designed to conduct in normal use.
All in all using a zener is better, its easier to get the right voltage, and the intention is clear (TVS diodes are used for protection from transient voltage spikes carrying lots of energy). But a TVS diode will work if you figure out the right voltage for the current it will be carrying (look at the curves in the datasheet).
Using a Zener, 415V - 120V = 295V on the EL84 screen. OK
But, with 450V B+, and an output transformer primary DCR drop of 20V, there must be 430V on the EL84 Plate . . . Ouch!
Did you mean EL34?
But, with 450V B+, and an output transformer primary DCR drop of 20V, there must be 430V on the EL84 Plate . . . Ouch!
Did you mean EL34?
Sorry I did not make it clear. My design is different from the original one.
In my design, I plan to use the triode/pentode tube 6F5P (ECL85). The pentode section has a maximum plate voltage at 300V and a maximum screen voltage at 250V. I want to triode strap the pentode section and set the plate voltage at 300V.
I am concerned about the high screen voltage, so I want to drop 20V or so from the plate voltage. One solution is to add a 1K grid resistor. This is what recommended for SV83 (6P15P).
The alternative is to use Zener diodes. Some people like them.
When I saw the TVS diodes bypassed with a capacitor were used in the commercial amplifier, I just want to know whether this is a better solution. If so, what part should I purchase?
Please let me know which one you would prefer from your past experiences.
In my design, I plan to use the triode/pentode tube 6F5P (ECL85). The pentode section has a maximum plate voltage at 300V and a maximum screen voltage at 250V. I want to triode strap the pentode section and set the plate voltage at 300V.
I am concerned about the high screen voltage, so I want to drop 20V or so from the plate voltage. One solution is to add a 1K grid resistor. This is what recommended for SV83 (6P15P).
The alternative is to use Zener diodes. Some people like them.
When I saw the TVS diodes bypassed with a capacitor were used in the commercial amplifier, I just want to know whether this is a better solution. If so, what part should I purchase?
Please let me know which one you would prefer from your past experiences.
I would just use a small screen resistor. I usually use 240R or 270R on power tubes, and 1k on smaller tubes like 6J1P.
I wouldn't worry about burning out the screen if you strap it as a triode. I run many tubes with a screen rated for 150V at 320V in triode. I've had very few issues.
If you really want to use a diode, I'd go zener for the same reasons Mark outlined.
I wouldn't worry about burning out the screen if you strap it as a triode. I run many tubes with a screen rated for 150V at 320V in triode. I've had very few issues.
If you really want to use a diode, I'd go zener for the same reasons Mark outlined.
If the screen current drops significantly, the zener will drop less voltage than its specified voltage rating (it will not be in the zener mode).
This can easily occur in push pull stages that are Class AB, they can have screen current drop to extremely low levels.
That might cause some transients as the zener voltage goes in and out of the zener region.
For single ended, this is less of a problem, because single ended is Class A, and the screen and plate do not go into cutoff
(very few want to listen to music that is clipped badly on their Hi Fi amplifier; Guitar amplifier - maybe OK)
Not for Ultra Linear circuits, but for those who are using Pentode/Beam Power Modes:
If the screen current will not be extremely high, a resistor in series with a VR tube/VR tubes in series will regulate the screen voltage (one of the simplest ways to regulate the screen voltage).
Depending on the design particulars, a zener might work very well.
Depending on the design particulars, a VR tube might work very well, but it dissipates more B+ power than the zener.
And as already mentioned, a small to medium resistance in series to the screen may be all that is needed for good performance.
This can easily occur in push pull stages that are Class AB, they can have screen current drop to extremely low levels.
That might cause some transients as the zener voltage goes in and out of the zener region.
For single ended, this is less of a problem, because single ended is Class A, and the screen and plate do not go into cutoff
(very few want to listen to music that is clipped badly on their Hi Fi amplifier; Guitar amplifier - maybe OK)
Not for Ultra Linear circuits, but for those who are using Pentode/Beam Power Modes:
If the screen current will not be extremely high, a resistor in series with a VR tube/VR tubes in series will regulate the screen voltage (one of the simplest ways to regulate the screen voltage).
Depending on the design particulars, a zener might work very well.
Depending on the design particulars, a VR tube might work very well, but it dissipates more B+ power than the zener.
And as already mentioned, a small to medium resistance in series to the screen may be all that is needed for good performance.
Last edited:
In the case of minimum zener current, just tie the zener to ground with a pull down resistor that will always pull enough current to make the zener work?
The power dissipated is high when screen current is significant , the TVS is much more robust , made to withstand high peak power .
I would not risk to use a Zener , can be very fragile and short easily
I would not risk to use a Zener , can be very fragile and short easily
Last edited:
That is assuming you will have 300 volts on the plate
Under load or with transformer voltage drop.
It could drop to 290/270
I really wouldn't worry about it and strap it in triode mode.
Or make life simple and easy and use a resistor around 600 to 800 ohms
Under load or with transformer voltage drop.
It could drop to 290/270
I really wouldn't worry about it and strap it in triode mode.
Or make life simple and easy and use a resistor around 600 to 800 ohms
Maximum ratings
Peak pulse power: 1.5 KWatt (10 x 1,000msec)
18.2 KWatt (8 x 20msec)
Steady state power dissipation: 3 Watts
from datasheet ...
If you know what is the role of such suppression diode , than it is obvious that is much stronger than a zener with equivalent steady power and higher used in series with G2 .
Last edited:
I see another issue with the 120V drop solution:
Lets suppose the supply voltage is 400V. The anode swing in the negative half cycle is approx. 350V. But the g2 swing will be 120V less, that is 230V. The result is reduced output power.
The proposed series resistor in g2 is mostly for preventing oscillation.
I had no issue with running a 807 tube at 425V triode strapped. The data sheet recommends g2 300V max.
Lets suppose the supply voltage is 400V. The anode swing in the negative half cycle is approx. 350V. But the g2 swing will be 120V less, that is 230V. The result is reduced output power.
The proposed series resistor in g2 is mostly for preventing oscillation.
I had no issue with running a 807 tube at 425V triode strapped. The data sheet recommends g2 300V max.
That is a comercial UL amplifier and 415V is too much for EL84 screen ...
Of course power is "lower" 😀 this way
A resistor can't protect the G2 , would be than to high in value for normal operation .
Of course power is "lower" 😀 this way
A resistor can't protect the G2 , would be than to high in value for normal operation .
Last edited:
The commercial amplifier is Kora Flash, 25 watts EL84M push pull. I believe Kora is no longer in business. the designer for the amp is Francois Philibert. Someone traced the circuit and posted here: HIFIDIY??-????????el84???? - Powered by Discuz!