Ca 21V ptp, this means 6-7W/8ohms. This with ECL86, DN2540 and 10k Raa.
Don´t think it is so simple to dial in anyway, as current-balance must be kept. Will do some empirical sims to try to find ut the optimal (added)CCS-current.
Don´t think it is so simple to dial in anyway, as current-balance must be kept. Will do some empirical sims to try to find ut the optimal (added)CCS-current.
21V (7.5W) sounds good work for ECL86. But with 6V trafo, the primary voltage will be >400V per side, if simple calc serves. We will need 600V or 800V FETs to keep transients out, I'd say.
I have found an IXYS 100V, 100mA, 25W depletion MOSFET at 0,67EUR each. Think I will have to order a bag, as these also have a lot lower Ciss than the Supertex.
900V? 10M90? I'd be interested to see if distortion changes at all vs the DN2540 in the Schaded configuration. Based on published curves, the devices gfs vs Id curves are somewhat different FWIW
For the original servo setup that oscillated, there were two cascadeded low pass filters, each setting a pole in the servo frequency response. One of them needs to be set at least 10X lower in freq. so as to act as a dominant pole to avoid oscillation.
Arrow has the IXTP01N100D listed, last time I checked.
Don
Arrow has the IXTP01N100D listed, last time I checked.
Don
At such small Idss and such large Vgs(off) is not very suited for use in a ccs.
Edit: but of course you weren't to use it in a ccs, so, never mind
Last edited:
Don, I was changing the antritriode's time constant (higher) when the circuit went bang. I am not confident that the servo transistors' pole could be set so low without resorting to FETs - which I was avoiding for accuracy reasons, and the need for even higher voltage margin between the current sense and the tail supply.
I began thinking that the CCS could be fed from a doubler out of the mains trafo, zener regulated, and use a darlington servo with very low pole to filter the signal out. But the parts count of that solution brought me around to the simplified circuit.
Lars, I may have to try some for that price!
Not sure about using it for antitriode, though. Won't it be full-ON when the B+ is applied? Make sure your CCS FET is a 600V part!
I like the DN2540 for specifying a max limit on temperature stability of threshold voltage, useful for our dc balance error budget.
For antitriode duty, I recommend the ST-microelectronic STP5NK90Z in this position: 900V 125W handling and low price.
Actually, the circuit shown above will withstand 600V on the CCS, and if you bolt the FET down to a big piece of chassis metal, you should get away with putting in a short-circuited output valve without blowing the fuse! Use a gate zener if your FET is not the protected type.
As you all bright boys probably understood the IXYS FET is 1000V, not 100V. Its the one jkeny mentioned.
Last edited:
Yes, Lars, they are also recommended by KevinC at K&K audio as sonically superior to DN2540N5 in cascoded ccs
I have some of those IXYS Fets coming - Rod, have you got yours yet & tried them? Anybody else built this?
At such small Idss and such large Vgs(off) is not very suited for use in a ccs.
Edit: but of course you weren't to use it in a ccs, so, never mind
Vgs(off) is -2.5 min, -5V max. I haven't hooked one up on the bench yet, but from the spec it's in the ballpark of a 10M45 (which is listed as "Vg"= -5V min and operates at -2.5 or so at lower currents).
High Vgs(off) is IMO a great attribute for CCS use because it provides for higher Vds on the bottom device in a simple cascode connection (less capacitance, higher current potential).
It's the low capacitance that makes me most happy
looks to me like an excellent part for cascode anode loads on driver tubes. The capacitance is about as good as the FQP150/160 parts.My biggest wish is for IXYS to provide comprehensive device data. Supertex (IR, Hitachi, etc) all provide lengthy data sheets but IXYS information content is scant indeed. Maybe you need to have a "relationship" with them to get the test data.
Michael
Hey jkeny,
Will you build with ECL86? If going EL84, another driver is better. Have a great solution for that one, tried by another member of this community.
Last edited:
I will be building with ECL86 currently in Baby Huey config which I'll be retaining so should be able to transfer the tubes between this anti-triode & the BH. Now, one question - the output stage imbalance is <1mA IIRC so it should be suitable for a toroid o/p transformer - any other considerations that might prevent this?
I'll have to bend Shoog's ear to suggest a suitable PS toroid to use.
The Telemas at about 80VA are great at the job. They are available from Farnell and RS spares.
I am currently working up to version two of the DC7 and will be ordering some for myself. I will also be using some 125V adjustable voltage regulators for the cathode CCS which are available from RS spares.
Shoog
I am currently working up to version two of the DC7 and will be ordering some for myself. I will also be using some 125V adjustable voltage regulators for the cathode CCS which are available from RS spares.
Shoog
Thanks Shoog, is "RS spares" a cheaper area of Radionics that I haven't seen? What secondary voltage, I can't remember. Do I need a centre tap on both primary & secondary. Are we talking 115-0-115: ?-0-?
Yes its Radionics. I don't find their prices excessive on things like transformers, if you can get it up to a certain level then postage is free.
You have to work back from the impedance you need to get the voltage.
(primary voltage/secondary voltage)^2*speaker impedance=primary impedance
So lets assume an 8ohm speaker
A 12V:115+115V will give you
(230/12)^2=367
367*8=2K9
So assuming you need an 8K lets try again
A 9V:115+115V will give
(230/9)^2*8=5K2
A 7V:115+115V will give
(230/7)^2*8=8k6
Which is near enough.
Shoog
You have to work back from the impedance you need to get the voltage.
(primary voltage/secondary voltage)^2*speaker impedance=primary impedance
So lets assume an 8ohm speaker
A 12V:115+115V will give you
(230/12)^2=367
367*8=2K9
So assuming you need an 8K lets try again
A 9V:115+115V will give
(230/9)^2*8=5K2
A 7V:115+115V will give
(230/7)^2*8=8k6
Which is near enough.
Shoog
Thanks again - they have 80VA 115-0-115:9-0-9 for 22 each - not bad for a tube o/p trafo 🙂 Are the secondary windings on the outside & can be unwound to 7V?
Edit: I'll PM you off-line so as not to contaminate this interesting thread 🙂
Edit: I'll PM you off-line so as not to contaminate this interesting thread 🙂
Last edited:
Hey Shoog,
You are not completely correct about the impedances. The formulas are right though.
You have to use the secondary voltage with NO load to calculate.
So when using a 30W 115+115 : 6+6V the secondary is 6,8V unloaded and the value used for calculations(9,2k).
I´d say a 7V is to high for the ECL86 if we assume 7V means more in reality. It is very easy to unwind the secondaries though, and I have done it quite a few times.
We used the 30W Nuvotem to the SEPTOR and it works great. So anything from 30-80VA will surely work.
You are not completely correct about the impedances. The formulas are right though.
You have to use the secondary voltage with NO load to calculate.
So when using a 30W 115+115 : 6+6V the secondary is 6,8V unloaded and the value used for calculations(9,2k).
I´d say a 7V is to high for the ECL86 if we assume 7V means more in reality. It is very easy to unwind the secondaries though, and I have done it quite a few times.
We used the 30W Nuvotem to the SEPTOR and it works great. So anything from 30-80VA will surely work.
- Status
- Not open for further replies.