It appears that there is no gate stopper on the fet. I believe that most of these fets have an internal zener gate to source, but it pays to make sure that the one you are using does - if not you should add it. (12 - 15V should be fine) The fet should fail more or less instantly if the zener diode is not present as remarked in a previous post.
A 100 ohm resistor very close to and in series with the gate should be added IMHO. You can cut the pcb etch right near the gate and solder in a 1/2W or (smaller resistor, even smd) right next to the fet footprint.
The -250V supply is a problem, you need to figure out why it is so high, might explain why the fets are failing. (More likely it is an indication of blown fet followers.)
A 100 ohm resistor very close to and in series with the gate should be added IMHO. You can cut the pcb etch right near the gate and solder in a 1/2W or (smaller resistor, even smd) right next to the fet footprint.
The -250V supply is a problem, you need to figure out why it is so high, might explain why the fets are failing. (More likely it is an indication of blown fet followers.)
Hi Kevin - The internal FET zeners are certainly able to protect the gate from ESD events, and minor transients.
But depending on them as protection from a 150V transient - that would appear if B+ is missing (if the B+ is shorted, or dc fuse is blown?) would not be wise, IMHO. From the data sheet of the specified FET 2SK2700:
(1) Gate-Source breakdown voltage: ±30V
(2) Leakage at Vgs = ±30V: 10µA
This zener protection is clearly not intended for large-signal gate protection, and now that I look it up, this is confirmed by TOSHIBA::
From their FET FAQ:
Q.: Is it possible to use the Zener diode between the gate and source for surge absorption?
A: The Zener diode between the gate and the source is intended to protect a MOSFET from electrostatic breakdown.
For surge absorption in the end-use application environment, add an external Zener diode.
FAQs of MOSFETs / Bipoler Transistors | TOSHIBA Semiconductor & Storage Products Company | Asia-Pacific
If there is no external protection on the PCB, a change is needed to maintain reliability of the FET: add a gate-source zener of 0.5W 9V1 to 15V; Cathode to gate.
I agree, if there is no gate stopper, one certainly needs to be inserted. SMD 1206 resistors are easy to work with - albeit with a headset magnifier, if the constructor is in my advanced age range.
Yes, -250V at the bias supply must not be ignored, and will increase the risk to the FETs.
But depending on them as protection from a 150V transient - that would appear if B+ is missing (if the B+ is shorted, or dc fuse is blown?) would not be wise, IMHO. From the data sheet of the specified FET 2SK2700:
(1) Gate-Source breakdown voltage: ±30V
(2) Leakage at Vgs = ±30V: 10µA
This zener protection is clearly not intended for large-signal gate protection, and now that I look it up, this is confirmed by TOSHIBA::
From their FET FAQ:
Q.: Is it possible to use the Zener diode between the gate and source for surge absorption?
A: The Zener diode between the gate and the source is intended to protect a MOSFET from electrostatic breakdown.
For surge absorption in the end-use application environment, add an external Zener diode.
FAQs of MOSFETs / Bipoler Transistors | TOSHIBA Semiconductor & Storage Products Company | Asia-Pacific
If there is no external protection on the PCB, a change is needed to maintain reliability of the FET: add a gate-source zener of 0.5W 9V1 to 15V; Cathode to gate.
I agree, if there is no gate stopper, one certainly needs to be inserted. SMD 1206 resistors are easy to work with - albeit with a headset magnifier, if the constructor is in my advanced age range.
Yes, -250V at the bias supply must not be ignored, and will increase the risk to the FETs.
Hy Rod,
I regularly use 2SK2700 (and 2SK3564, AOT1N60, or new SiC C2M1000170D) as source follower with +400..420V and (via source resistor) -200...220V without any trouble. I never had gate stopper there, nonetheless are never oscillate.
But... never use single (Ixys or Supertex) MOSFET as current source!
Cascode is safer (and better at any respect) with protection zener from "lower" FETs gate to source.
Even so I lost many 400-450V FETs due to the power-on transients in my test applications. The final layouts are safer, but now I never use this medium voltage FETs over 300V.
Kilovolt FETs (for example IXTP01N100D) are more robust and typical survivors. 🙂
I regularly use 2SK2700 (and 2SK3564, AOT1N60, or new SiC C2M1000170D) as source follower with +400..420V and (via source resistor) -200...220V without any trouble. I never had gate stopper there, nonetheless are never oscillate.
But... never use single (Ixys or Supertex) MOSFET as current source!
Cascode is safer (and better at any respect) with protection zener from "lower" FETs gate to source.
Even so I lost many 400-450V FETs due to the power-on transients in my test applications. The final layouts are safer, but now I never use this medium voltage FETs over 300V.
Kilovolt FETs (for example IXTP01N100D) are more robust and typical survivors. 🙂
Hi Bela,
Yes, if the B+ is present the circuit will usually start up OK without bursting the FET gate. But we should always try to make the circuit robust! The circuit in the diagram will certainly give trouble if the bias voltage comes up first - or if some unexpected transient comes along.
With gate stoppers and oscillation, the problem depends on the layout, but I don't like to take a risk for the cost of one small resistor.
I agree - when possible use kilovolt-rated FETs - they are worth the small extra cost.
Yes, if the B+ is present the circuit will usually start up OK without bursting the FET gate. But we should always try to make the circuit robust! The circuit in the diagram will certainly give trouble if the bias voltage comes up first - or if some unexpected transient comes along.
With gate stoppers and oscillation, the problem depends on the layout, but I don't like to take a risk for the cost of one small resistor.
I agree - when possible use kilovolt-rated FETs - they are worth the small extra cost.
The power supply uses a 5AR4 rectifier tube. This tube comes up slowly. So the -250V comes up quickly and then it takes a little while for B+ voltage to come up. So there is a while that we have -250V without the B+ voltage.
In my practice most of B+ supply "delayed" (switched), so B- is stable (30-60s), before B+ available.
Source follower FET "mortality" rate in my amplifiers is ZERO!
IMHO the weakness is the 10M45 FET. Red plate of the (VAS) driver tube is possible if the current source FET is shorted.
Is this tube remains good? Cathode voltage? Cathode-heater, grid-cathode shorting...
The other possible reason of failure is in the "boutique" components. I always measure all resistors (and sometimes capacitors too) before installation.
If you remove coupling capacitor between first stage and source follower, both (first tube and source follower-output tube) stage adjustable and measurable separately.
If both stages operating points (voltages, currents) are right, install coupling capacitor. If nothing has changed, the amplifier DC stability is OK.
Source follower FET "mortality" rate in my amplifiers is ZERO!
IMHO the weakness is the 10M45 FET. Red plate of the (VAS) driver tube is possible if the current source FET is shorted.
Is this tube remains good? Cathode voltage? Cathode-heater, grid-cathode shorting...
The other possible reason of failure is in the "boutique" components. I always measure all resistors (and sometimes capacitors too) before installation.
If you remove coupling capacitor between first stage and source follower, both (first tube and source follower-output tube) stage adjustable and measurable separately.
If both stages operating points (voltages, currents) are right, install coupling capacitor. If nothing has changed, the amplifier DC stability is OK.
Last edited:
I've had brds fail with stock bill of material parts, and boutique parts. The boutique resistors are all 1/4 watts Vishay naked Z foils. All larger wattage resistors are custom made Vishay Naked S foils, made by Texas Components. All .1% tolerance. I measured all resistors before installing, and they are right on. And I measure all my caps to, before installation.
I just went out to measure B- voltage. It went up to neg. 360 volts, then dropped to neg. 325 volts when B+ came up. Measured C6, and C7. C7 measured 100uf as soon as I put the meter on it. C6 read 50uf and slowly climbed to 100uf. Maybe C6 is damaged? Anyway, I'm sending to Mouser for the stock Caps for these spots. Lets see if this fix's the B- problem.
The zeners are really needed. The circuit may be working, thanks to the internal ESD protection, but it makes no sense to rely on them against that level of overvoltage.
78-BZX55C12-TAP is the Mouser stock-code for a suitable diode. They're a few cents each.
Does it start up with -360V right on the 300B grid? What voltages on C6 and C7?
Building the amp with the parts that the designer intended is a good plan. Sometimes "boutique" can have unintended properties.
78-BZX55C12-TAP is the Mouser stock-code for a suitable diode. They're a few cents each.
Does it start up with -360V right on the 300B grid? What voltages on C6 and C7?
Building the amp with the parts that the designer intended is a good plan. Sometimes "boutique" can have unintended properties.
- Status
- Not open for further replies.