anatech said:
Even if Firewire standard can provide some good currents
I think anatech advice here is a good one!
This is what M-Audio writes in User Manual
of their top of line Audio soundcard with Firewire:
See also my attached Screen capture from the PDF Manual.
Regards, lineup
Attachments
anatech said:
No.
Actually I have only a Creative Soundblaster Live! 24-bit, so far.
I am, as I have told in other topics
been thinking of order this one:
Terratec Aureon 7.1 Universe
Details:
http://sounden.terratec.net/modules.php?op=modload&name=News&file=article&sid=197
It has not got Firewire, but has got RCA and full digital in/out 24/96.
Also Phono RIAA preamp input.
An externally hosted image should be here but it was not working when we last tested it.
But I am ckecking for good alternatives.
this one looks like a good laternative for HiFi Audio quality.
M-Audio FireWire AudioPhile
Details:
http://www.m-audio.com/products/en_us/FirewireAudiophile-main.html
An externally hosted image should be here but it was not working when we last tested it.
=========================================
Actually, I ordered Terratec Universe in november
but they had it not in store.
And I got tired of waiting (6-7weeks!) and cancelled my order.
No hurrry.
I may research the market of PC soundcards for Hifi a bit more
and see if there are even other good alternatives.
I might buy a PC Card Oscilloscope.
But both of those cards in this post
have extremely advanced software included in package!
This in combination with high standard 24-bit digital in/out
can go a long way, for testing audio signals.
I think.
lineup
Hi lineup, Shawn,
I picked up a Creative Blaster X-FI extreme music. It will sample at 96 KHz 24 Bit also, just two channel. They run about $120 CDN and mount in a PCI slot.
See here for specs.
Very good audio quality for just listening on top of that.
-Chris
I picked up a Creative Blaster X-FI extreme music. It will sample at 96 KHz 24 Bit also, just two channel. They run about $120 CDN and mount in a PCI slot.
See here for specs.
Very good audio quality for just listening on top of that.
-Chris
Toshiba 2SC2240
I purchased a bunch of Toshiba 2SC2240's and here is how they measured up today:
hfe - Qty
357 - 3
345 - 3
337 - 7
333 - 5
330 - 3
320 - 4
315 - 5
311 - 7
308 - 7
296 - 8
292 - 4
287 - 1
281 - 5
273 - 3
257 - 2
250 - 5
210 - 1
207 - 1
204 - 1
Just wondering what you think? If I match sets of these devices, or any other for that matter, what tolerance should I allow if any? Different tollerances for outputs versus input stage devices? I now understand it depends on the application and circuit but general advice and experience is nice to hear.
Thanks,
Shawn.
I purchased a bunch of Toshiba 2SC2240's and here is how they measured up today:
hfe - Qty
357 - 3
345 - 3
337 - 7
333 - 5
330 - 3
320 - 4
315 - 5
311 - 7
308 - 7
296 - 8
292 - 4
287 - 1
281 - 5
273 - 3
257 - 2
250 - 5
210 - 1
207 - 1
204 - 1
Just wondering what you think? If I match sets of these devices, or any other for that matter, what tolerance should I allow if any? Different tollerances for outputs versus input stage devices? I now understand it depends on the application and circuit but general advice and experience is nice to hear.
Thanks,
Shawn.
for input pairs
for example 2 NPN I would match for close Voltage B-E
And I would try to use the pair that are close in measure
with the highest gain for input stage
Vbe matching and HFE matching of transistors from same batch
is practically same thing.
Those with similar gain value usually have almost same Vbe, usually in region of 0.60-0.70V.
if you have a complementary input, with 4 transistors
= 2 NPN + 2 PNP you try to find 2 close
If building amplifier with a longtailed pair NPN input
and using those with values in your posts:
I would start with those with highest gain = 3 x 357
Those two of these 3, that have best Volt Base-Emitter match I would use.
Gain=357 and gain=345 is almost same.
The last digit is not all the world.
So you can could just as well measure Vbe of
those SIX transistors with highest gain, and pick 2 of them for input.
For other stage, like VAS it is good with high gain,
and closer to output stage it is good with match,
but most important is good match and precision in input stage.
Why?
Because here is where most often the feedback and control of output take place.
So a slight mismatch in other stages, are corrected, error-correction,
in this input stage.
I am talking here about the normal amplifiers, with negative feedback from output signal
-- back to input stage.
Here output is compared by the input signal by the 2 transistors in this pair.
If they are mismatch, which stage will correct them?
----------------------
There are different amplifiers, more or less different from this 'normal'
and so many factors can decide where you need to match.
----------------------
So called fully symmetrical amplifiers are among the most difficult to match.
This is because the lack of good complementary matching NPN-PNP transistors.
Especially difficult is N-FET and P-FET as well as N-MOSFET and P-MOSFET.
lineup
for example 2 NPN I would match for close Voltage B-E
And I would try to use the pair that are close in measure
with the highest gain for input stage
Vbe matching and HFE matching of transistors from same batch
is practically same thing.
Those with similar gain value usually have almost same Vbe, usually in region of 0.60-0.70V.
if you have a complementary input, with 4 transistors
= 2 NPN + 2 PNP you try to find 2 close
If building amplifier with a longtailed pair NPN input
and using those with values in your posts:
I would start with those with highest gain = 3 x 357
Those two of these 3, that have best Volt Base-Emitter match I would use.
Gain=357 and gain=345 is almost same.
The last digit is not all the world.
So you can could just as well measure Vbe of
those SIX transistors with highest gain, and pick 2 of them for input.
For other stage, like VAS it is good with high gain,
and closer to output stage it is good with match,
but most important is good match and precision in input stage.
Why?
Because here is where most often the feedback and control of output take place.
So a slight mismatch in other stages, are corrected, error-correction,
in this input stage.
I am talking here about the normal amplifiers, with negative feedback from output signal
-- back to input stage.
Here output is compared by the input signal by the 2 transistors in this pair.
If they are mismatch, which stage will correct them?
----------------------
There are different amplifiers, more or less different from this 'normal'
and so many factors can decide where you need to match.
----------------------
So called fully symmetrical amplifiers are among the most difficult to match.
This is because the lack of good complementary matching NPN-PNP transistors.
Especially difficult is N-FET and P-FET as well as N-MOSFET and P-MOSFET.
lineup
Lineup,Mr. Nuts & Bolts man , your comments seem sound to me. I never read anyone ever say, "Vbe matching and HFE matching of transistors from same batch
is practically same thing. Those with similar gain value usually have almost same Vbe, usually in region of 0.60-0.70V.
This is a profound statement to me which I never took note of before. Now I will do more testing of Vbe to screen the devices further. If what you say is true, then one test "may" be just as good as the other. Perhaps not in all cases but in most and this is what I am interested in.
I would like to match the devices and sort and collect them in sets for future projects. I could have purchased four 2SA970 & four 2SC2240 for $1.50 each ($12.00) but I chose to purchase 75 pcs each for $0.20 each ($30.00).
I will measure the hfe of the complimentary device 2SA970 in the next day or so and post the results. Then I will set up for Vbe!
Cheers,
Shawn.
I never read anyone ever say, "Vbe matching and HFE matching of transistors from same batchis practically same thing. Those with similar gain value usually have almost same Vbe, usually in region of 0.60-0.70V.
This is a profound statement to me which I never took note of before. Now I will do more testing of Vbe to screen the devices further. If what you say is true, then one test "may" be just as good as the other. Perhaps not in all cases but in most and this is what I am interested in.
I would like to match the devices and sort and collect them in sets for future projects. I could have purchased four 2SA970 & four 2SC2240 for $1.50 each ($12.00) but I chose to purchase 75 pcs each for $0.20 each ($30.00).
I will measure the hfe of the complimentary device 2SA970 in the next day or so and post the results. Then I will set up for Vbe!
Cheers,
Shawn.
CBS240 said:In your opinion, if you had matched pairs of descrete j-fets, would you use the ones with higher measured Vgs or the lower Vgs for the input pair? or does it even matter. Might one likely be more linear than the other, or maybe slightly different 'linear' operating area
The one with higher Vgs is more liniar if it is operated at higher Ids curents
Hi,
now that you have low gain, medium gain and high gain groupings, select according to Vbe at the operating current.
Try to match Vbe to within 5mV and keep the gain with 5%.
If you can get both then great, if you cannot get pairs with both, then put more priority on Vbe matching.
A good amplifier design should be impervious to gain variations. But you are looking for better than good. Hand selected matching comes into excellent category. Keep in mind temperature and testing accuracy. I suggest that once you have found matching Vbe in a gain group you retest hFE, they will have changed.
Is this for a dual complementary front end?
Then deliberately mismatch the NPN from the PNP. If you match then the two LTPs become very sensitive to output offset error with changes in temperature of the four input junctions. A 10% to 20% difference between NPN & PNP makes a near fixed input offset current that you trim out later and then the output offset is less sensitive to junction temperature.
now that you have low gain, medium gain and high gain groupings, select according to Vbe at the operating current.
Try to match Vbe to within 5mV and keep the gain with 5%.
If you can get both then great, if you cannot get pairs with both, then put more priority on Vbe matching.
A good amplifier design should be impervious to gain variations. But you are looking for better than good. Hand selected matching comes into excellent category. Keep in mind temperature and testing accuracy. I suggest that once you have found matching Vbe in a gain group you retest hFE, they will have changed.
Is this for a dual complementary front end?
Then deliberately mismatch the NPN from the PNP. If you match then the two LTPs become very sensitive to output offset error with changes in temperature of the four input junctions. A 10% to 20% difference between NPN & PNP makes a near fixed input offset current that you trim out later and then the output offset is less sensitive to junction temperature.
TomWaits said:Lineup,Mr. Nuts & Bolts man ,
your comments seem sound to me.
I never read anyone ever say,
"Vbe matching and HFE matching of transistors from same batch
is practically same thing. Those with similar gain value usually have almost same Vbe, usually in region of 0.60-0.70V.
This is a profound statement to me which I never took note of before. Now I will do more testing of Vbe to screen the devices further. If what you say is true, then one test "may" be just as good as the other. Perhaps not in all cases but in most and this is what I am interested in.
.
I look forward to see your results:
vbe vs. hfe
I wont say it is the same. But almost.
In my experience, those from same 100 package, or so
those with more similar Vbe value, also will be somewhat having the same level of gain.
Another thing I would add, in how to match a good NPN pair ( or PNP )
is
if possible make your matching test
at same current and same collector-emitter voltage
that you will use in your amplifier.
This will give ( P = U x I ) same effect (milliwatt)
and so the temperature in transistors will be close to the temperature when used in the amplifier.
lineup
Hi Andrew,
I don't often disagree with you, but I've found that matching for hFE is the best and produces the least distortion. Matching the NPN and PNP pairs (comp. diff pairs) offers improved sound quality and the least, lowest drift in an amplifier.
Hey Shawn,
Where did you buy your transistors? I'm getting low myself.
-Chris
I don't often disagree with you, but I've found that matching for hFE is the best and produces the least distortion. Matching the NPN and PNP pairs (comp. diff pairs) offers improved sound quality and the least, lowest drift in an amplifier.
Hey Shawn,
Where did you buy your transistors? I'm getting low myself.
-Chris
AndrewT said:
Cool! Your words make me feel like the extra effort is worth it.
lineup said:
Can you guys help me build a circuit to screen the devices for the front end of the Krell Clone KSA 100 mkII? This is the amp I am going to select parts for right now.
anatech said:
They came from a virtual unknown (to me) Audio Lab of Ga . They charged $12.00 US for shipping but the postage paid on the package was $4.00! Perhaps the $12 covers handling? It’s ok. They took a few weeks to show up as well. They seem to have an interesting medley of components in stock. I would buy from them again.
Cheers,
Shawn.
Attachments
TomWaits said:
Due to the use of ZENER diode D1, 39V
we can guess that voltage across those 4 input transistors will be like 32 volt collector-emitter
Have a look at my attached NPN transistor tester.
Instead of 30.6V as shown, you can use 32.6V from your lab supply (one adjustable LM317 will do this!)
From this circuit we can measure it all, by only using Volt meter.
The base-emitter voltage of BC550C is like 0.650-0.700V
and this voltage will be across
R1 (470 ohm ) - the current set resistor
------------------------------------
1. C - E = Vce
2. B - E = Vbe
3. (E - G)/R1 = Ib + Ice = current in transistor
4. (A - B)/10k = Ib = base current
5. HFE = Ice/Ib
------------------------------------
I set the current to ~1.5mA (~0.700V/470),
as I think each pair has got current source like 3 mA. (39V/12K1, R14)
With 50-50% = 1.5mA + 1.5mA sharing
For measuring PNP pair, 2SA970,
you just mirror my circuit and use one BC560C for set current.
C1, 1nF is added for stability. Avoid oscillations.
lineup
Attachments
Re: Toshiba 2SC2240
Toshiba 2SA970 here is how these measured up:
hfe - Qty
346 - 1
355 - 1
366 - 1
383 - 8
388 - 5
398 - 1
405 - 1
408 - 2
415 - 5
418 - 3
422 - 10
427 - 3
433 - 4
436 - 2
439 - 3
444 - 2
452 - 9
463 - 3
465 - 2
491 - 2
Fakes! I hope not but it wouldn't be the end of the world to me.
Tough little transistors to find in stock. Check 'em out. I can get some through a franchised disti but it may take a while.
Shawn.
Toshiba 2SA970 here is how these measured up:
hfe - Qty
346 - 1
355 - 1
366 - 1
383 - 8
388 - 5
398 - 1
405 - 1
408 - 2
415 - 5
418 - 3
422 - 10
427 - 3
433 - 4
436 - 2
439 - 3
444 - 2
452 - 9
463 - 3
465 - 2
491 - 2
Fakes! I hope not but it wouldn't be the end of the world to me.
Tough little transistors to find in stock. Check 'em out. I can get some through a franchised disti but it may take a while.Shawn.
Attachments
I'll try to fire it up on the weekend and continue measuring these devices.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.