DN2540

[merlin el mago]

Thks (grazie 1000) Gianluca.

 

[sheep]

How does a pair of transistors cascoded like the Bottlehead C4S perform?
It uses 2 red LEDs and 2 transistors in a cascode.

 

[SY]

It would appear to be pretty much the same as the diyAudio CCS boards offered a few years ago and the CCS shown in Morgan Jones's "Valve Amplifiers." Performance is very good- not as good as cascoded DN2540, but good enough that it won't be the limitation in any circuit you're likely to build.

 

[sheep]

I'll try them. They should be better than LM317 CCS.

 

[digits]

Hey ok I got mine in, if you still need a few.

 

[tmblack]

How do you calculated the resistor values for the DN2540 current source?

 

[SY]

You can approximate them from the published curves, but there's enough batch-to-batch variation that you might want to set up a test jig to determine the resistor value (IME, the resistor value will be pretty consistent device-to-device in the same batch). For 10mA, a good starting point will be about 300R. For 20mA, about 120R. For 5mA, about 560R.

 

[tmblack]

Are they good as current sources in the range of 0.5mA to 2mA?
Some say a gate resistor should be used, or can I just ground the gate to the source resistor?

 

[SY]

You absolutely want a gate-stopper resistor. Anything from 300R-3k will work, but you want the resistor body as close to the gate as possible. The work very well at low currents, but I would definitely cascode them.

 

[Magz]

A single 10M45S is better than a single DN2540. But a cascaded DN2540 beats a 10M45S big time. Not sure if you can cascade 10M45S.

Check this out: CCS performance measurments


Damn. Wish I had know that. My last order from Mouser was for 73 euro's and I had to pay 20USD shipping.

K and K audio says the IXYS IXTP01N100D is better than the DN2540. Unfortunately. Europeans are NOT allowed to buy that part from the US of A.

Yes, you can cascade the IXYS part; I use the 10M45S cascaded in my preamp as a plate load, with SiC Schottky diode bias. It sounds excellent.

 

[SY]

The cascoded DN2540 is insanely better than you'd ever need (>100M impedance, noise in the pA range). I don't know if the IXYS part is any better, but you're already well past the point of diminishing returns, so use whichever you can get easiest/cheapest.

 

[Sheldon]

They work very well at low currents, but I would definitely cascode them.

You have a wide choice of JFET's for the bottom device in the cascode at low currents, cause the DN2540 does the heavy voltage lifting.

Sheldon

 

[SY]

Absolutely, but you have to pick a bottom FET with a lowish pinch off voltage. Fortunately, the DNs are only a buck or so, so it hardly seems worthwhile to use something else.

 

[Alain Poitras]

Hi all,

I know this is not exactly the right forum, but I'm searching for some DN2540 depletion mosfets, and it seems I cannot find them here in Italy; so, I know some of you have already used them, and hope you can help me with a shop address here in Europe (i would prefer to do not buy them oversea).
As an alternative you could give me some equivalent depletion mosfet easier to be found here in Europe.
Thanks in advance

Ciao,
Giovanni

I buy two on eBay from UK about 3 months ago, I paid 8,30$ CAN (£5,21 GBP) shipping included.

But I don't try them yet, it is for my winter experimental circuits.

Alain.

 

[Sheldon]

Absolutely, but you have to pick a bottom FET with a lowish pinch off voltage. Fortunately, the DNs are only a buck or so, so it hardly seems worthwhile to use something else.

I used 2SK170 for on one project, just because I had only a couple of the DN's and had some of the SK's sitting around to pair them up with.

Sheldon

 

[SY]

Nothing better than free!

 

[Bandersnatch]

hey-Hey!!!,
On the cascoding I think it would be useful to run one of the IXYS bits on top. They have about double the gate-source voltage and would allow more d-s voltage for the lower one to work in. The larger voltage is ov primary concern cause gate-srain capacitance drops so rapidly at these d-s voltages. This would allow use of the higher than 400V parts here too( to do the 'heavy lifting' ).
cheers,
Douglas

 

[tmblack]

Using the DN2540 cascode as an active load I sill get hum through, but not if 2SK170 is used. What gives?

 

[SY]

Which circuit are you using?

 

[tmblack]

I tested the DN2540 with the modified JFET BOZ that uses 2SK170 in a Common source configuration.
Using another 2SK170 as a CS did reduce hum but not a DN2540 single or cascoded CS.