yeah sorry about that
as for the 2sk369-v i bought 4 quads matched tightly with all 16 falling within 0.2ma and all over 20ma.
for sure i'll let people know if all turns out kosher
Last edited:
Thanks Patrick
but when you suggested 2sk369 i thought you had something else in mind...
V grade is 14 to 30 mA that is why i asked .
In the article, it was recommended to use in each SEN circuit a quad set of 2SK170BL with Idss 8mA. This was meant to be used with AD1865 which has output current of 2mA.
The article also recommend that you increase the bias if your DAC outputs more current, as e.g. in the case of ES9018. So just for the high current output alone (whether 9018 or else), you should choose the V grade (e.g. 2SK170V).
The reason to choose 2SK369V was to lower Zin (seems to be more critical for ES9018). 2SK369 has double the Yfs of 2SK170 and hence Zin can be halved, roughly.
Again, all described in the article.
Patrick
The article also recommend that you increase the bias if your DAC outputs more current, as e.g. in the case of ES9018. So just for the high current output alone (whether 9018 or else), you should choose the V grade (e.g. 2SK170V).
The reason to choose 2SK369V was to lower Zin (seems to be more critical for ES9018). 2SK369 has double the Yfs of 2SK170 and hence Zin can be halved, roughly.
Again, all described in the article.
Patrick
> Who's selling?
Give me a couple of days.
If I can get some V grade at reasonable price, I'll get them matched to quads and make them available to the Vref subscribers.
I think you would need 4 sets of quads per 2 channels of balanced SEN.
So 200 pcs would supply 10 subscribers, assuming 25% matching rejects for quads.
But see if I can get them first ....
Patrick
Give me a couple of days.
If I can get some V grade at reasonable price, I'll get them matched to quads and make them available to the Vref subscribers.
I think you would need 4 sets of quads per 2 channels of balanced SEN.
So 200 pcs would supply 10 subscribers, assuming 25% matching rejects for quads.
But see if I can get them first ....
Patrick
here
i was able to get the set for 70aud, so cheaper than listed. not cheap, but well matched and i needed to get them on the way to help out. now of course Patrick will sell them for half that
I took it as a reasonable deal considering the quality of the match and the relative difficulty finding them, but i'm sure i paid more than needed
Last edited:
Thanks Patrick,
I do have the article ( book ) but should have read with more attetion i guess
.From the point of view of Yfs the 2sk372V is even better, it has triple typical vale of the s2k170V, it does however have half the power so this should be considered.
Last edited:
How would this circuit look like when the dac has balanced current output?
Will two CCS be needed (one for each phase) and if yes, may they share the same supply
The article in LA is great, but it does not describe how to mate the CEN/SEN with DAC's "featuring" current/voltage off-set. This is left to those "skilled in the art" and has been worked out after publication in this thread.
I'm afraid that I'm "skilled in other arts", so advice for dummies would be appreciated
Cheers,
Nic
P.S. The referred to circuit is shown in post #182 and described in post #170.
Thanks Patrick!
> How would this circuit look like when the dac has balanced current output?
In this particular circuit, there is no interaction between the two sides of the balanced circuit.
So you just use each phase of the balanced output of one channel with a separate SEN.
i.e. You need one PCB for each balanced channel.
> Will two CCS be needed (one for each phase) and if yes, may they share the same supply
Yes.
And they absolutely may not share the same supply.
The supply swings together with the output signal.
So the supply of the positive output would be 180° out of phase with that for the negative output.
Or, 4x 9V batteries for each balanced channel.
Nic,
As I have already explained, I cannot posted a solution for a circuit that I have not tested.
I know how it should work, but someone has to test and report before you should build.
Jeremy has kindly promised to do.
So please be patent for a while and let him show you once he gets it going.
My apologies,
Patrick
In this particular circuit, there is no interaction between the two sides of the balanced circuit.
So you just use each phase of the balanced output of one channel with a separate SEN.
i.e. You need one PCB for each balanced channel.
> Will two CCS be needed (one for each phase) and if yes, may they share the same supply
Yes.
And they absolutely may not share the same supply.
The supply swings together with the output signal.
So the supply of the positive output would be 180° out of phase with that for the negative output.
Or, 4x 9V batteries for each balanced channel.
Nic,
As I have already explained, I cannot posted a solution for a circuit that I have not tested.
I know how it should work, but someone has to test and report before you should build.
Jeremy has kindly promised to do.
So please be patent for a while and let him show you once he gets it going.
My apologies,
Patrick
yeah sorry for the delay chaps, cant be helped while i wait for postage and my own business is pretty busy this time of year. i'll get on it when everything arrives. i got my DK order sent to a friend in the us to ship with other things and this caused longer than anticipated delay due to how busy he is at the moment. besides my 2sk369-v havent got hear yet either
cool, i'll probably grab some spares. so i got tracking info from my friend, sent fedex international priority at great expense, partially due to the delay caused here and partially because i need things in it too. jesus they know how to charge, cost nearly 400 dollars. it will all be here monday, hopefully the 2sk369-v will be here before then as well.
So for those who intend to use the Vref version for the likes of ES9018, we recommend using a quad set of 2SK369V for each single channel of SEN, or 4 sets of quads for 2 channel balanced.
We bought 250 pieces of 2SK369V (as shown in post #511), then curve traced 12 of them at random, then tested a pair of them in circuit to make sure there is no question at all with them. We have now measured and sorted them all. Out of 250 pieces, we can only make up 8 sets of 4 quads (i.e. each set has 16 FETs for 2 channel balanced SEN) to the level of match that we would use ourselves. The matching loss rate is then almost 50%.
Here is a typical set (see attached). Within each quad, Idss under 9V Vds (= SEN operating conditions) is within 0.2mA. This is only important for equal heat dissipation, and is equivalent to <+/-1% dissipation difference. But when paired as recommended, the different in bias between top & bottom pairs are below 0.02mA. This is important to input DC offset. And 0.02mA is equivalent to ~0.4mV input offset.
The 8 sets will only be offered to Vref kit subscribers. Please indicate interest at the GB thread :
http://www.diyaudio.com/forums/group-buys/196585-sen-cen-all-jfet-iv-converter-evaluation-pack.html
As we are only interested in enabling you to build the circuit, they will be offered at well below commercial rates :
http://www.diyaudio.com/forums/swap-meet/156408-matched-2sk369-v-sale.html
... just like we did with the 2SJ74BL's in the original GB. Price to be announced later.
If the demand far exceeds supply, we will consider ordering another 250 for further matching.
Patrick
We bought 250 pieces of 2SK369V (as shown in post #511), then curve traced 12 of them at random, then tested a pair of them in circuit to make sure there is no question at all with them. We have now measured and sorted them all. Out of 250 pieces, we can only make up 8 sets of 4 quads (i.e. each set has 16 FETs for 2 channel balanced SEN) to the level of match that we would use ourselves. The matching loss rate is then almost 50%.
Here is a typical set (see attached). Within each quad, Idss under 9V Vds (= SEN operating conditions) is within 0.2mA. This is only important for equal heat dissipation, and is equivalent to <+/-1% dissipation difference. But when paired as recommended, the different in bias between top & bottom pairs are below 0.02mA. This is important to input DC offset. And 0.02mA is equivalent to ~0.4mV input offset.
The 8 sets will only be offered to Vref kit subscribers. Please indicate interest at the GB thread :
http://www.diyaudio.com/forums/group-buys/196585-sen-cen-all-jfet-iv-converter-evaluation-pack.html
As we are only interested in enabling you to build the circuit, they will be offered at well below commercial rates :
http://www.diyaudio.com/forums/swap-meet/156408-matched-2sk369-v-sale.html
... just like we did with the 2SJ74BL's in the original GB. Price to be announced later.
If the demand far exceeds supply, we will consider ordering another 250 for further matching.
Patrick
Attachments
Last edited:
Patrick, although the concern with dissipation is a good one, have you had a chance to see how close to the zero temperature coefficient point these parts are, typically? I've found 170BL to be often quite close to a zero drain current tempco at 0 Vgs. In fact some of them require a slight forward bias for the zero point! The canonical approximate operating point for JFETs is about 600mV away from the pinchoff voltage, based on the simple models.
When I was measuring some BF862 parts for various parameters recently, I found that they too are very stable with temperature at 0 Vgs. A nice attribute.
When I was measuring some BF862 parts for various parameters recently, I found that they too are very stable with temperature at 0 Vgs. A nice attribute.
Last edited:
While it is an interesting point, it is of no issue here as long as the upper and lower JFETs track each other thermally, as is the case.
This thermal concept allows you a wider choice of devices at varying operating points.
It alleviates the need to operate at zero tempco.
If the FET is close to zero tempco by coincidence, so much better.
Patrick
This thermal concept allows you a wider choice of devices at varying operating points.
It alleviates the need to operate at zero tempco.
If the FET is close to zero tempco by coincidence, so much better.
Patrick
I had a question in mind about the CEN, since I read the article: I remember, that in your analysis of the F5, you point out that the 2SK170 and 2SJ074 are not true complementary, and suggested an asymmetric circuit to compensate for that.
Why in this topology it doesn't apply ? (Or is negligible)
Thanks,
Davide
Why in this topology it doesn't apply ? (Or is negligible)
Thanks,
Davide
It applies also here, but as explained already in the article, Kirchoff's law guarantees i_in = i_out.
Whether there is equal current sharing between N & P JFET does not matter at all.
You have to bear this in mind -- it is not an amplifier. It is a current conveyor.
It just suffles current from input to output, while keeping Zin loa and allowing a high R_load.
Patrick
Whether there is equal current sharing between N & P JFET does not matter at all.
You have to bear this in mind -- it is not an amplifier. It is a current conveyor.
It just suffles current from input to output, while keeping Zin loa and allowing a high R_load.
Patrick
