| Onvinyl |
Hi there,
I'm new to this place, but hopefully in the right place...
As you might guess, I'm mostly into vinyl reproduced music.
I'm currently trying to build a mc-stage for my diy-ed phonostage 'le pacific', a two-stage jfet Design using the KISS-approach developed by a french audiophile.
It basically works; but while trying out several jfets, there is some strange behavior. It might be a kind of oscillating, I don't know. It produces a frequent and stable popping. Playing with load resistor,, it is possible to 'silence it', but the signal is still modulated with it.
The stage consists basically of a 4K Load-Resistor, a 2SK170 JFET (tried 3,5 and 6,3 mA IDSS), tried 0-68R Sourceresistor, and 0,65uF Cap at the output. Driven with 24V batterie, and this stage is etrxa de-coupled with 10.000uF Elna electrolyte.
Used with 2SK369 the stage works (with and without electrolytes), but lacks a bit of bass and tightness.
Where can I start to track the problem down? I'm more or less experienced with existing designs, but new to try some of my own.
greets,
Rudiger |
|
|
| Onvinyl |
I might add, it's a breadboard design, and the MM-part works fine even without housing or shielding.
The MM-part is as follows: |
|
|
| NickC |
you may need to decouple the stages from one another. circuit are known to oscillate without decoupling. normally go with a RC to form a lowpass filter. I don't have the proper formally for the proper values though. Or try a LC lowpass filter(better no voltage drop). C is normally 220uf-470uf , R is normally low 10 ohm.
Try searching for power decoupling tutorials, you will learn alot from it. Also used a good cap for c (BLackGAte,ELna, Muse etc). You used a 10uf Elna only that is actually bypass only. Doesn't totally isolate the stages from one another( sk170 at input and the sk170 at the output). Decoupling is much better. For bypass try blackgate high q nx work like magic but cost like hell. Just my two cents
P.s batteries supply has high impedance , use caps to bring it down. High speed caps is a must here |
|
|
| HDTVman |
Hi Onvinyl
NickC is on track with his advice. Are you adding another gain stage to what you have now to give you the extra gain you need for the MC pickup?
If so decoupling the power supply between stages and bypass caps should make a big difference. Bypassing needs to be done locally at the gain stage for it to do any good and every gain stage needs it in low level, high gain amps.
You may want to think about using pcbs with a good ground plane. This may also help. Others may think differently about this but it has always worked well for me.
Good luck
Bruce |
|
|
| Jocko Homo |
Try them sometime. Then thank me.
Jocko |
|
|
| dice45 |
Jocko,
have to agree, my one FET MC headamp also went better with a grid stopper oops gate stopper.
BTW, longtimenosee! :)
Onvinyl,
spotted your QQ on vinyl_lebt, thought you were in good hands asking hifidaddy/analogguedaddy :)
My one FET single-ended MC-headamp used 12V supplied by AA NiCd batteries, it used one 2SK43 per channel, 2SK240 also is usable. drain resistor was 1k, input load resitir also was 1k, both tantalum oxide. Output coupling cap was 0.5 µFSiemens MKV 25839.
PS caps were huge Siemens electrolytics, 100.000. µF bypassed with 2x1nF silver mica.
If you use the search engine, you may even find a pix i posted.
Gate stopper was added later so i remember, nonmagnetic SMD, 100Ohm. Or was it 50? |
|
|
| Onvinyl |
Hi,
in my case, (correct) decoupling was the way to go. Found two 47R Resistors and soldered them in.
@HDTVman:
It's a three-stage design now. The last two are as described in the circuit, the first one as described in my first post.
What do you all mean by 'gate stopper resistor'? 100 Ohms in series? This is already the case.
@NickC:
What are 'high speed caps'?
Bernhard, found your pic of your mc-prepre. These are some decent electrolytes....
The whole thing is a breadboard design by now. I will not use a PCB finally, rather p2p as Bernhard did it.
althoug my design works now, sonically I should get quite more out of it.
Thank you for your great suggestions,
Rudiger |
|
|
| dice45 |
Onvinyl,
email me if you need some micas, Might have some left.
Oh and please try out those MKVs for coupling caps. They are magic. witchcraft.
Unfortunately uncheap. Bürklin seels them. |
|
|
| Onvinyl |
Hi Bernhard,
got my Coupling-C's from steinmusic, shouldn't be too bad...
For the mica's: well, I'll check out and let you know.
Currently, I'm looking for good 'soldering stripes' and thinking about a good (wood) housing, probably with a coppper shielding. Any hints here?
Rudiger |
|
|
| NickC |
The phono amp looks easy to build for mm. Could i use sk369 instead of sk170 as i can order them from vacummstate? or where can i order sk170 matched?
Hows the sound of the phono amp? |
|
|
| Onvinyl |
Hi NickC,
I tried those first in my mc-stage. They work, but lack bass and tightness. That is the experience of hifidaddy as well.
I wouldn't use them. This phonoamp is supposed to be a quite good one, really high grade.
My current design does not sound bad, though it is not a killer by now.
There are suggestions to try 2SK240 double-fet instead.
Rudiger |
|
|
| NickC |
| how would you connect the circuit then? new to fets world (bjt user). Also some tips on preventing esd would be good? thanks |
|
|
| peranders |
| quote: | Originally posted by NickC
new to fets world (bjt user). Also some tips on preventing esd would be good? thanks |
Every semiconductor are sensitive but some are more....
JFET's and BJT will normally not create much trouble but MOSFET's can be sensitive especially those with no gate protection what so ever. Small signal MOSFET's like BS170 are not protected. Also power fet's. Exception from that are Hitachis lateral types 2SJ50, 2SK135 and newer models. If you aren't sure, consult the datasheets. |
|
|
| NickC |
| anybody knows where to get the jfets 2sk170 GR model? and how to matched them.(circuit design would be great) |
|
|
| Jocko Homo |
Yes, but only if you want blues. 2SK147s, blue and probably violet.
Jocko |
|
|
| dice45 |
Rüdiger,
best hint for ceramic soldering strips i can give you: lay your hand on an old (non-functional) Tektronix oscilloscope and carefully dismantle, slaughter, butcher it. It is full of those strips.
Circuits: best avoid the side cutter at all and do it with your soldering iron.
Before i forget it: do it with respect and let the mastership of the Tek people how to do proper p2p wiring slowly soak in. Zen style :)
Use silver solder, it is needed for those Tek cermic soldering strips. |
|
|
| NickC |
is it possible to use more jfets at the begining of the circuit. it reduces more noise than using one jfet rite. but with have to add resistance to the source for equally distribution of load rite?
what value should it be? and at the output stage should more jfets be used (increase gain) similar to pass pearl. what are the benefits of my suggestion and disadvantages too |
|
|
| Morello |
I suggest that You implement a RF-filter at the input.
2k ohms in series and 100pF shunt to ground.
I have designed a RIAA-stage comprising one 2SK170 and one OPA2134 per channel. It works excellent with extremely low noisefloor.
Best regards\Morello |
|
|
| Werner |
Does anyone have a 2sk240 datasheet? Seems to be a bit hard to find on the net.
Thanks,
W |
|
|
| jcarr |
Werner: I am pretty certain that the 2SK240 has been out of production for some years now, and this is likely why you cannot find a datasheet. From memory, this was a 40V medium-gm low-noise N device. The complementary was called the 2SJ75, I think.
Although I don't think I have a full data sheet, I probably have at least some basic numbers, so if you cannot find anything else, I'll go through my parts catalogs.
regards, jonathan carr |
|
|
| Werner |
Jonathan,
thank you very much. I have bought a number of NOS 240s, and all I really need is the correct pinout. Even though G is probably the center pin, and D and S are mostly interchangeable, I'd rather verify this before I attempt to blow up anything.
Thanks again,
W |
|
|
| hifidaddy |
| quote: | Originally posted by Werner
Jonathan,
thank you very much. I have bought a number of NOS 240s, and all I really need is the correct pinout. Even though G is probably the center pin, and D and S are mostly interchangeable, I'd rather verify this before I attempt to blow up anything.
W |
Hi Werner,
verified !
in the datasheet, it says that the top of the aluminium case has a drain marker, and when looking at what is inside (just two normal TO92 JFETS, mounted back to back) it is clear they must be interchangeable ...
well, the 2SK240 is exactly a double 2SK170, and has the same data.
regards,
Hartmut |
|
|
| Werner |
OK. Thanks Hartmut.
W |
|
|
| jcarr |
Dear Werner, I am always happy to help.
Turn the device over, so that you are looking at the bottom (pin side). From left to right, top row: drain 1, gate 1, source 1.
From left to right, bottom row: drain 2, gate 2, source 2.
This is from the yearly FET data book published by CQ Shuppan, which is something of a Japanese industry standard.
>D and S are mostly interchangeable.<
It appears that Toshiba thinks so, too. The 2SK240 was one of those Toshiba oddities in which, instead of putting two wafers into a hermetic can, they simply took two identical TO-92 plastic devices and put them into an aluminum can, with the flat surfaces touching. If the two TO-92 devices were identical, you would _expect_ that the pin orientation would be drain 1, gate 1, source 1, source 2, gate 2, drain 2. But no, the drains and sources exit together. So either the two TO-92 devices are _not_ identical, or the differences between source and drain don't matter very much on these Toshiba JFETs.
I wonder how many other dual JFETs this is true for (even if the manufacturer doesn't say so). Because if so, it would certainly make it easier to create PCB layouts for differential amp stages. Anyone tried this for, say, the Vishay-Siliconix 2N5564~5566 family?
regards, jonathan carr |
|
|
| Werner |
Jonathan,
FWIW, the datasheet of the single BF245 mentions explicitly that source and drain can be interchanged.
W |
|
|
| Onvinyl |
Hi again,
I'm up and about to give my circuit a proper housing and all.
How do I treat the lead acid batteries right? They spend a voltage of +24V, and at the moment I'm feeding them through a simple PSU mainly constisting of a L200 current+voltage regulator wich provides 26,4 V. Normally, i don't disconnect batteries from PSU while listening. Is this the way to go? Or should I charge batteries with 27,6V and disconnect, when charged? Or design a tricky circuit that dis- and reconnects depending on state of batterie?
Advice is very appreciated,
Rudiger |
|
|
|
