Bo Hansen DI layout

[Whoops]

This may be stating the obvious but you'll have a hard time getting 2SK170 unless you have some already.

Hi Newmarket,
the 2SK170 is quite easy to find:

http://www.banzaimusic.com/2SK170BL.html
http://www.banzaimusic.com/2SK170GR.html

I think I have quite a few in my stock at the workshop

Can't wait to try this out with the LL1538XL on DoubleBass Piezo of an artist I work with.
I was never satisfied with the sound with 1M input DI's, only when we used a Jensen DI that you can select 10M for the input

 

[Bo Hansen]

Gus,

I absolutely agree with you.

A Piezo pickup without impedance converter sitting very near the pickup must have a short low cap. cable, if it should feed a input with 10 Mohm or more.

So I hope Whoops has thought to have this JFET modified DI-box very close to the Piezo pickup, otherwise it will not be good at all.

However, if there is an "active" impedance converter inside the instrument where Piezo pickup is mounted, the DI box should have a normal instrument input impedance around 1 Mohm as most guitar amps have,  same as my DI box design in original version.

--Bo

 

[Whoops]

A Piezo pickup without impedance converter sitting very near the pickup must have a short low cap. cable, if it should feed a input with 10 Mohm or more.

So I hope Whoops has thought to have this JFET modified DI-box very close to the Piezo pickup, otherwise it will not be good at all.

It's nice to have a 10M input DI because I can't control the equiment of the musicians I work with (Live) , or that might work in a a specific gig.
A lot of musicians just have a piezo without any impedance converter.

With a 2 to 3 meters low capacitance cable, I don't mind the minor HF reduction, it's preferelable to me than an impedance mismatch as that affects both ends of the spectrum in a nasty way.

 

[Whoops]

I'm wondering why the GR version of the 2SK170 is much more expensive than the BL....

2SK170BL  €0,96

2SK170GR    €4,76

 

[Bo Hansen]

Idss selection for 2SK170 is GR 2,6-6,5mA, BL 6-12mA, V 10-20mA.

I have also noticed that GR is the most expensive, think the reason is that there are the most left of BL and V on the market.

Buy BL it will be perfect, Idss value have no major importance in this connection

But watch out when you buy transistors from China, it sells many fake types there.
2SK170 was only a Toshiba manufacturing, so if you have a genuine Toshiba sample you can see how the text looks, it can be difficult, but if it's very different you must be aware.

--Bo

 

[Newmarket]

But watch out when you buy transistors from China, it sells many fake types there.
2SK170 was only a Toshiba manufacturing, so if you have a genuine Toshiba sample you can see how the text looks, it can be difficult, but if it's very different you must be aware.

--Bo

Well I was wondering about those 2SK170. I see Banzai are, from their website at least, based in Germany. NOS ?  What you think ?

 

[Whoops]

Well I was wondering about those 2SK170. I see Banzai are, from their website at least, based in Germany. NOS ?  What you think ?

Bo is right, if you go on Ebay there's a lot of 2SK170 option from China, most or all of those are probably fakes.

I use as a supplier Banzai a lot (for at least 10 years) and I never have any problem with fake components,
but I was also alert to the fact that they are the only suplier I can find stocking 2SK170BL and GR.

I checked my stock and I have 2x 2SK170BL from Banzai,
I did measurments of both with my tester unit and also took some pictures,
what do you think?

 

[Bo Hansen]

Newmarket,

Hard to say, because there is no real picture of them.

You can email them and ask if they can guarantee that it's genuine Toshiba.

Here are two generations that I know are genuine Toshiba, because they are delivered in sealed factory packs.

 

[Bo Hansen]

 

[Whoops]

another Source:

https://www.bdent.com/2sk170bl-toshiba-audio-transistor.html

 

[Whoops]

From this thread:

http://www.diyaudio.com/forums/pass-labs/188484-2sk170bl-2sj74bl-2.html

 

[Bo Hansen]

Guys,

Do not spend much work to finding true 2SK170,  search instead for the "old work horses" 2N3819 or 2N5457 from Motorola/ON-semi, National, Fairshild, Siliconix.
Or the  European BF245A from Philips and some others.
It's almost available from all component companies.

Just make sure you get the pin-out right, especially 2N3819 which is manufacturer-dependent.
Also change the 100k resistor to around 130k as I mentioned earlier.

These will work very well, it's just a simple source follower.

--Bo

 

[Bo Deadly]

I checked my stock and I have 2x 2SK170BL from Banzai,
I did measurments of both with my tester unit and also took some pictures,
what do you think?

If the current listed is Idss, then I'd say they're definitely fake. Not even close. Transconductance is -2*Idss/Vp which should be like 20. But based on these pics it's only like 3.

I have 2SK170 from various sources. Some from fairly convincing vendors who had pictures of the original Toshiba packaging. Others were from China with dubious listing details. After a detailed spectral analysis [1], I concluded that the China ones were indeed fake but only because the gain vs noise characteristics were notably different from the multiple other more consistant sources (4.3dB more gain without an increase in the noise floor). Meaning the Chinese fakes were pretty damn good. So I kept them but labelled the bag "2SK170 good fake" accordingly. I have actually used them in circuits and they are perfectly good.

However, the transconductance was never less than 18 and the one's that I believed to be genuine were as high as 29. Yours are less than 3.

Having said that, note that for this application, it will be almost totally unimportant. Just about any JFET will do as this circuit. It just needs to bias properly (which may not be the case if the Idss is really only 1mA according to that meter).

[1] https://groupdiy.com/index.php?topic=66423.0

 

[Bo Hansen]

Squarewave,

Many thanks.
Very interesting evaluation and measurements in the attached link.

--Bo

 

[Whoops]

If the current listed is Idss, then I'd say they're definitely fake.

Having said that, note that for this application, it will be almost totally unimportant. Just about any JFET will do as this circuit. It just needs to bias properly (which may not be the case if the Idss is really only 1mA according to that meter).

Might be the meter as I don't know how well this $15 meter measure JFETS

Will measure them with my Peak Atlas DCA75 just out of curiosity, I just dont have the unit with me at the moment

Thanks

 

[Gus]

You could add something like this to the DI
http://www.till.com/articles/PreampCable/
Have another jack with a tip ring sleeve to connect to the jfet cable
I believe Analogguru posted a jfet cable buffer on the web in the past, however I have not found it.
You could build a circuit like the follower Schoeps uses in a remote capsule like you see on the CMC schematic

 

[Whoops]

You could add something like this to the DI
http://www.till.com/articles/PreampCable/
Have another jack with a tip ring sleeve to connect to the jfet cable
I believe Analogguru posted a jfet cable buffer on the web in the past, however I have not found it.
You could build a circuit like the follower Schoeps uses in a remote capsule like you see on the CMC schematic

A DI with a 10Mega input was what I needed and asked Bo if it was possible on his DI.
Thanks Bo for the input and advices, problem solved

thanks

 

[Bo Hansen]

Whoops,

No problem, you are welcome.

Btw.
If you want to experiment with a small JFET impedance converter near the Piezo pickup, to remove the cable's disadvantages as previously discussed here, you can use this simple design and only change  the 1 Mohm resistor to 10 Mohm or higher.

Then you can feed my DI box input as original version.

But I think you're quite happy with the modification you've made now, especially if you're using a short low cap. good cable between the double bass and the DI box.

--Bo

 

[Whoops]

Hi thanks Bo,
it's a good idea also, but I will build quite a few of your DI boxes, and it's nice to have same with the 10 Meg input.

The DI's will be used mainly in live sound, a DI with the 10M input will be good also for non-piezo, line level signals and electric guitar/bass are quite happy with it also.

A lot of DI's have more than 1M input impedance for the same reasons,
here are just a few examples:

Leon Audio Active DI Box Mk2b (30Mega)
http://www.leonaudio.com.au/active.htm

Countryman Type 85 (10Mega)
http://www.countryman.com/type-85-direct-box/

Radial PZ-DI (10Meg, 1M, 220M)
http://www.radialeng.com/product/pz-di/

Radial StageBug SB-4  (5Meg)
http://www.radialeng.com/product/stagebug-sb4/

Avalon U5 (3Mega)
http://www.avalondesign.com/instrudi.html

 

[Bo Hansen]

Whoops,

Yes I know, most active DI boxes on the market are around 10 Mohm.

However, the most used input impedance for most of all guitar and bass amps is 1 Mohm, which means that pickups and volume/tone controls on passive instruments have a behavior as musicians are more used to, compared with a much higher impedance.
Ok, the difference is not big but it is still there.

Btw,
My own DI boxes I designed in recent years have actually fixed 10 Mohm input impedance on the input for passive instruments, and they are also JFET both on the input and output amp, with a more advanced circuit.

So I do not live as I learn 

--Bo