yes please do that ...everybody will love the smoke !
Rugged yes but duty is different and protection also ...said before
Rugged yes but duty is different and protection also ...said before
Oh and by the way
square wave testing ...or extended bandwidth tests above 40KHZ have nothing to do with what you listen
Theoretical approach will require any amplifier to perform equally 20-20KHZ marketing approach of one amplifier that performs 200KHZ or more is obviously useless for Human hearing .
Square wave tests on inductive load have to do with stability and speed and will present you phenomena like ringing , clip behavior, overshot or undershot and so on...( i left oscillation out of the picture just for being polite )
Things like that cannot be seen in a typical sine wave of 1khz and resistive load ...
That on its own shows how much you know about amplifiers ....
square wave testing ...or extended bandwidth tests above 40KHZ have nothing to do with what you listen
Theoretical approach will require any amplifier to perform equally 20-20KHZ marketing approach of one amplifier that performs 200KHZ or more is obviously useless for Human hearing .
Square wave tests on inductive load have to do with stability and speed and will present you phenomena like ringing , clip behavior, overshot or undershot and so on...( i left oscillation out of the picture just for being polite )
Things like that cannot be seen in a typical sine wave of 1khz and resistive load ...
That on its own shows how much you know about amplifiers ....
Last edited:
Here is newark's copy of the BDW93/94 datasheet from fairchild:
http://www.farnell.com/datasheets/269747.pdf?_ga=1.145191794.775858257.1414767027
No Ft listed, but Cob typically is 140 pf @ 1 v & 50 pf @ 100 v
The BVces test current of 100 ma kind of suggests not using even the C grade at more than 40 v.
I can't believe these darlingtons including the TIP142/147 are so cheap, $.92 to $1.15, when 2SC5200 are $2.44 and 2SA1943 are $2.59. Last year the 2S parts were >$4.
http://www.farnell.com/datasheets/269747.pdf?_ga=1.145191794.775858257.1414767027
No Ft listed, but Cob typically is 140 pf @ 1 v & 50 pf @ 100 v
The BVces test current of 100 ma kind of suggests not using even the C grade at more than 40 v.
I can't believe these darlingtons including the TIP142/147 are so cheap, $.92 to $1.15, when 2SC5200 are $2.44 and 2SA1943 are $2.59. Last year the 2S parts were >$4.
For instance or example Indianjo this specific semi from Fairchild is a TO220 device while the original device from Philips that originally produced these parts ( i might be wrong about this ) was a TO3P device
There is much confusion about this but it doesn't stop there
Please check alldatsheet.com for any of the products of any family
you will clearly notice that there is resistors between driver and output transistor that vary depending on the manufacturer and in real life obviously this will vary farther more depending on the production , the specific order, batch date of production and so on
Such a mess around them simply make them unsuitable for any design that aims to performance and stability .
Also ...let us not forget that both ft and cob is useless to be included in a datasheet when the title is Solenoid driver for low switching application ( tip 142-147)
Anyone today is willing to work with that ?
There is much confusion about this but it doesn't stop there
Please check alldatsheet.com for any of the products of any family
you will clearly notice that there is resistors between driver and output transistor that vary depending on the manufacturer and in real life obviously this will vary farther more depending on the production , the specific order, batch date of production and so on
Such a mess around them simply make them unsuitable for any design that aims to performance and stability .
Also ...let us not forget that both ft and cob is useless to be included in a datasheet when the title is Solenoid driver for low switching application ( tip 142-147)
Anyone today is willing to work with that ?
I won't be buying BDW93/94 since TIP142/147 are available in TO3P for $.20 more. But Andrew wanted the link to the datasheet from somebody better than alldata.
The reason I bought the TIP142/147 is specifically darlingtons were recommended by NatSemi for the LME49810 driver IC, which I bought 25 of. But I don't have any 4 ohm speakers that sound good, so TIP142/147 is not going to put out the 300 W the IC datasheet advertises on only +-40V. The IC pins are so close together I haven't figured out how to solder wires to the IC anyway, and I won't be buying any PCB (or anything) from C****.
Everybody that used the LME49810 is obsessed with thermal trak output transistors, which are **** expensive.
The five parallel TIP142 are making nice voltage regulators driven by a simple 1.3W zener diodes stack. Two parallel TIP147 worked great for the linear voltage regulator in my Schober Recital Organ amp too, that had blown about 5 germanium zener sense transistors in its life (built 1976) to look at the fried remains in the parts box. That amp with PNP germanium output transistors is now now my reliability king, playing records frequently while I try to work the bugs out of AX6 driver board for the dynakit ST120 chassis.
The reason I bought the TIP142/147 is specifically darlingtons were recommended by NatSemi for the LME49810 driver IC, which I bought 25 of. But I don't have any 4 ohm speakers that sound good, so TIP142/147 is not going to put out the 300 W the IC datasheet advertises on only +-40V. The IC pins are so close together I haven't figured out how to solder wires to the IC anyway, and I won't be buying any PCB (or anything) from C****.
Everybody that used the LME49810 is obsessed with thermal trak output transistors, which are **** expensive.
The five parallel TIP142 are making nice voltage regulators driven by a simple 1.3W zener diodes stack. Two parallel TIP147 worked great for the linear voltage regulator in my Schober Recital Organ amp too, that had blown about 5 germanium zener sense transistors in its life (built 1976) to look at the fried remains in the parts box. That amp with PNP germanium output transistors is now now my reliability king, playing records frequently while I try to work the bugs out of AX6 driver board for the dynakit ST120 chassis.
Last edited:
Yes, square wave testing is useful for the reasons you mention. I'm not about to drive a 100W square wave into any load I have because I don't have a load which can handle it. And I'm not about to push my amplifier over its limits.
In no way are the Darlingtons going to do >100W, and are a stretch at 75W, but for music they're fine at 75W, and in a limited bandwidth amplifier, 90W is going strong.
I'll post my Darlington amp doing full swing square waves into 8ohm (30W or so).
My guitar amplifier is 50W, but I've used overrated components because it's designed to be used in tough conditions and to deal with overdrive, etc. Hifi amp does high dynamic range music.
lets see first if you ever gona make a decent square at 10KHZ (even less will do )
lets this as a behavior in a combined load and will be enough even if it is in 5W ....
lets this as a behavior in a combined load and will be enough even if it is in 5W ....
About not having a load.
My amps have deteriorated enough over the years, I have sound owning a pair of 225 W 10 ohm sliding tap resistors very useful. When the wattage is not what is advertised, time to replace some parts again. P=V^2/R. V is measured with an ordinary analog VOM.
I re-e-capped the ST70 4 times in 46 years and output tubes were replaced thrice. 17.5 was the desired voltage out of that amp. When it sounded "polite" it was often putting out 7.
I caught an Allen S100 amp that the tech expert on organforum said "never needs e-caps in his experience", putting out 2 VAC on the speaker last year. New e-caps alone pulled it up to 15.
yeah, the Ohmmite ones are inductive, not pure resistive. So what- so are speakers.
BTW I don't have a working oscilloscope, curve tracer, gain meter, or capacitance meter below 10 UF. But I can repair amps.
My amps have deteriorated enough over the years, I have sound owning a pair of 225 W 10 ohm sliding tap resistors very useful. When the wattage is not what is advertised, time to replace some parts again. P=V^2/R. V is measured with an ordinary analog VOM.
I re-e-capped the ST70 4 times in 46 years and output tubes were replaced thrice. 17.5 was the desired voltage out of that amp. When it sounded "polite" it was often putting out 7.
I caught an Allen S100 amp that the tech expert on organforum said "never needs e-caps in his experience", putting out 2 VAC on the speaker last year. New e-caps alone pulled it up to 15.
yeah, the Ohmmite ones are inductive, not pure resistive. So what- so are speakers.
BTW I don't have a working oscilloscope, curve tracer, gain meter, or capacitance meter below 10 UF. But I can repair amps.
Last edited:
It's such a mission finding decent transistors here in SA! I don't trust one place because I got fakes there before. Then there are a couple of places, but to find stock of what you want is mission impossible! I did find some transistors that I like. Has anyone got any experience with them? The outputs look good, but the drivers I have no idea about:
Drivers:
2SD669 / 2SB649 120V 1.5A 20W 140MHz TO-126
Outputs:
2SC3281 / 2SA1302 200V 15A 150W 10MHz TO-247
The thing is that I've used TIP35C/TIP36C a lot, and they work like a bomb, but they're pushing it at only 100V Vce, and they're getting expensive too. Those transistors I've listed are available for fairly cheap, and will justify massively rugged design.
For 135 W, I'd use two pairs on the output, and 4 pairs for 270 W. Would that be reasonable? Then I can get rid of the Darlingtons, because Hfe is enough that I can design the VAS to operate at low enough power.
While I'm at it, does anyone have a spice model for those above? I haven't done a search yet, but if they can be added here, it might be useful to more than just me.
Drivers:
2SD669 / 2SB649 120V 1.5A 20W 140MHz TO-126
Outputs:
2SC3281 / 2SA1302 200V 15A 150W 10MHz TO-247
The thing is that I've used TIP35C/TIP36C a lot, and they work like a bomb, but they're pushing it at only 100V Vce, and they're getting expensive too. Those transistors I've listed are available for fairly cheap, and will justify massively rugged design.
For 135 W, I'd use two pairs on the output, and 4 pairs for 270 W. Would that be reasonable? Then I can get rid of the Darlingtons, because Hfe is enough that I can design the VAS to operate at low enough power.
While I'm at it, does anyone have a spice model for those above? I haven't done a search yet, but if they can be added here, it might be useful to more than just me.
Ok, I see the 2SC5200 / 2SA1943 are now available for quite cheap here. When I started this thread they were difficult to come by and quite expensive. Now they're ok.
2SC3281 / 2SA1302 and 2SC5200 / 2SA1943
These are likely to be fakes, especially if they are Ebay cheapies.
The 2SD669 / 2SB649 120V 1.5A 20W 140MHz TO-126 have long been obsolete.
There are a few copies still available and if they actually meet the published specification, then they are very good drivers. Few other BJTs can match that published specification.
How do you test to see if they comply?
These are likely to be fakes, especially if they are Ebay cheapies.
The 2SD669 / 2SB649 120V 1.5A 20W 140MHz TO-126 have long been obsolete.
There are a few copies still available and if they actually meet the published specification, then they are very good drivers. Few other BJTs can match that published specification.
How do you test to see if they comply?
well ...
sorry to say things like that but you force me to
There is no chance that any one that know or deals with audio amplifier to work with tip35 and/or put them on any comparison with 1302-3281 or 1943 5200
I think that last time i used tip 35 or similar i was 14 ....
Also having fake transistors is not a question of a retailer ... I mean retailers buy and sell codes and numbers Hardly ever they deal with the make batch quality and so on Their main goal is to buy cheap and sell expensive if possible
Its our job to find out which of the transistors in the stock of the retailer are fake or not ...
sorry to say things like that but you force me to
There is no chance that any one that know or deals with audio amplifier to work with tip35 and/or put them on any comparison with 1302-3281 or 1943 5200
I think that last time i used tip 35 or similar i was 14 ....
Also having fake transistors is not a question of a retailer ... I mean retailers buy and sell codes and numbers Hardly ever they deal with the make batch quality and so on Their main goal is to buy cheap and sell expensive if possible
Its our job to find out which of the transistors in the stock of the retailer are fake or not ...
RS Components has 2SC5200 / 2SA1943, and the Fairchild ones they list look good. I think I've bought some before.
I think those drivers will be difficult to come by anyway. The one shop has them, but future stock will be an issue.
Rather than those, the best alternative I can find here is MJE15030/1.
Not sure what you mean by comply? Fake vs genuine? I'm not in the position at the moment, but I'd buy 9 NPN's and 9 PNP's, and inspect and break open one of each I guess?
I think those drivers will be difficult to come by anyway. The one shop has them, but future stock will be an issue.
Rather than those, the best alternative I can find here is MJE15030/1.
Not sure what you mean by comply? Fake vs genuine? I'm not in the position at the moment, but I'd buy 9 NPN's and 9 PNP's, and inspect and break open one of each I guess?
TIP35 don't compare. They were my transistor of choice 10 years ago because they were available and affordable. Now they're only in use in my guitar amplifiers, but I want to find a permanent replacement with better specs for audio amplification.
The problem with fakes is that if the retailer sells you a bunch of them, it's not easy to trust that retailer again. I try to buy from shops with a store-front, so that I can inspect the transistors first, but the two I can go to are generally very expensive and don't stock one or the other of the complementary pair. RS components have been a good retailer so far.
I said the 649/669 are available as copies.
The datasheets for the copies seem to be the same as the originals.
Do the copies actually achieve the performance of the originals?
That's what should be checked, but how?
Fakes are a different problem.
Fakes have the same package and label as the original, but are guaranteed to be different inside.
Ok, well that's fine. Whether they comply, I don't have the equipment.
On the other hand, the amplifier can be modified to use a MOSFET output. I really like the IRFP240/IRFP9240, but what are the limits? Is it simply the drain current possible at a certain case temperature? I know BJT's, but I have no experience with MOSFET output stages. See attached; I would think it can do the 120 W in 8 ohms with +-53 V supply, but can't be loaded with 4 ohms at that supply? I would expect it can do 130 W in 4 ohms with +-42 V supply, provided there is enough (and realistic) heatsinking?
And with those transistors, I'm guessing thermal feedback is also necessary?
Attachments
You NEED very good thermal tracking, try an "amplified Vbe" sensor transistor and either bolt it to the heatsing if a flatpack (TO220, etc.) or drill a small hole in the heatsink and insert it there with some thermal grease if TO92.
Forget that Led + resistor biasing, it will burn your amp as soon as it warms up, switching type MosFets have a high *positive* thermal coefficient.
Forget that Led + resistor biasing, it will burn your amp as soon as it warms up, switching type MosFets have a high *positive* thermal coefficient.
Two considerations, if I may:
1) Driving the HexFETs directly from VAS does not work well because of the HexFETs' high, non-linear input capacitance (no problem with laterals, but a problem with HexFETs). Solution - push-pull EF stage between the VAS and the HexFETs.
2) As JMFahey just mentioned - thermal feedback is a must. Well-tested configuration for HexFETs with EF drivers - MJE340 (TO-126) with the red LED in series with the emitter - Tempco is particularly suitable for (IRFP2x0).
1) Driving the HexFETs directly from VAS does not work well because of the HexFETs' high, non-linear input capacitance (no problem with laterals, but a problem with HexFETs). Solution - push-pull EF stage between the VAS and the HexFETs.
2) As JMFahey just mentioned - thermal feedback is a must. Well-tested configuration for HexFETs with EF drivers - MJE340 (TO-126) with the red LED in series with the emitter - Tempco is particularly suitable for (IRFP2x0).
- Status
- Not open for further replies.