As far as I'm aware, the present designs around the LT4320 (and others) can't be used at the higher rails of these amps - there was mention of higher voltage versions by "tvc" awhile ago but nothing available yet.
A rectifier bridge well worth mentioning that works really well are the Shindengen bridges
A rectifier bridge well worth mentioning that works really well are the Shindengen bridges
You know what? If you replace the old electrolytic capacitors, things will improve. Now, if you match the two input pairs (and to each other as close as you reasonably can), the distortion will really improve a lot. Subjectively it will sound much, much better as well. I also try to match the complimentary drivers because in this simple design, that makes a difference too. I think that this is the best improvement you can make using the original boards. Subjectively this should also get you most of the way to the performance of the new PCB designed by Bob.
Hey Dave,
I would be interested in testing a complete unit with Bob's new boards. Then, if the parts haven't been matched, maybe I should try one of these with matched pairs (which I would do of course). I can match BJTs within 1% using the jig for that purpose. This always makes an audible and measurable difference in my experience.
-Chris
Hey Dave,
I would be interested in testing a complete unit with Bob's new boards. Then, if the parts haven't been matched, maybe I should try one of these with matched pairs (which I would do of course). I can match BJTs within 1% using the jig for that purpose. This always makes an audible and measurable difference in my experience.
-Chris
Rails of up to +/-72VDC can be used with the LT4320 if there is a chip dedicated to each rail. However this requires that the two rails do not share the center tap, i.e. there are 2 secondaries each with its own 2 wires (4 secondary wires out of the transformer.) That is why the stock transformers in the DH-200, 220, 230 will not work with the 4320. There is a high voltage "ideal" rectifier circuit that has been designed, tested and is soon to be available for purchase which waits only for a Mouser shipment of appropriately spec'd mosfets. The mosfets are due for delivery in early February, and the HV rectifier will ship soon after. I believe this one will handle better than 200 VDC, but will not typcially work accross series secondaries with a center tap. This is because secondaries used in series must be in-phase..
Ron
Peter,
I have a couple XL-280's I'm in the processing of modding, as well as several DH-200's, a couple DH-500's and a P500. For an "original" XL-280 I would consider:
- Replacing the electrolytic caps on the input board. If original these caps are 30+ years old, and could likely need replacing (sounds like you are already doing this)
- Additionally you could consider upgrading some of the other caps. Though not likely to have gone "bad", there may be higher quality ones now available which were not available in the late 80's
- Replace and/or upgrade PSU main caps (currently 4 x 7.8k uF). You can find ones now with twice the capacitance which fit in the same space. Look for quality low-ESR, high ripple caps
- Consider some of the PSU design improvements presented by Bob Cordell in chapter 19 of his new book:
- Addition of high-quality bypass caps across reservoir caps including a 100uF electrolytic + lower value film caps
- Addition of Zobel network
- Add small resistance between capacitor banks to create RC filter which can substantially decrease ripple amplitude
- The higher voltage, lower current secondary on the XL-280 transformer feeds the input and VAS stages of the circuit. This secondary voltage is rectified via 4 x 1N4003 discrete diodes. One could consider replacing these with a set of faster, soft-recovery type diodes like Vishay HEXFRED, (Cordell chapter 19, section 5).
Generally speaking one of the most exciting possibilities I see at the moment for upgrading vintage amps, (while I wait for the release of the Cordell boards), is the implementation of an "ideal rectifier" using the LT4320 chip. A lot of effort and money can be spent getting rid of noise generated by the standard bridge rectifier, but all this noise can be eliminated at the source by using the LT4320. However, in the case of the XL280 the 4320 cannot be used without changing the transformer. I may consider doing this depending on the results I get from the other mods I plan to do (as time permits).
It’s great to have amps on a platform that so many years later is shared with others. A good basic physical design in these amps allows that.
I understand all your suggestions, and will do many of them when the time comes. Thanks for taking the time to outline them.
I’ve not heard of the LT4320, will look into that. I would think additional reservoir decoupling would better be done at input of driver board, rather than at bulk caps, but that’s a minor item.
Couple of questions...
Have you placed an NTC thermistor on AC to limit surge? What part/value did you use?
What part number is a good quality binding post for speaker outputs? ...the stock ones are annoying.
I’d like to know what you use to test and match transistors? Should pairs also be thermally coupled?
As an aside, something I generally don’t see in modification suggestions is the replacement of the whole bulk power section with rectified line voltage driving two independent switchers, or just two ac input switchers.
This could give you massive power reserve with essentially dual mono topology. Caps would be dramatically smaller for less ripple, and that ripple frequency would be way above audio band.
I used to design switchers, back in the day when 100 kHz was the norm (these days it’s much higher). We were very concerned with noise, as these switchers drove codec linecards for telephony (idle channel noise a big yield problem). So we regulated the secondary of the switcher with magnetic amplifiers (small high permeability toroids)...which have the added benefit of forcing all rectifiers into soft reverse recovery...almost zero noise.
Eventually, I discovered that we could make the noise irrelevant by synchronizing the switching freq to the codec framing pulse, but we had already managed to noise to low levels enough with magamps.
(A mag amp is a small toroid directly in output rectifier path with a control winding on it, such that during the off time of the rectifier, you can preset the remanence point of the magnetic. This allows you pulse width control at the next on cycle. Because it’s a series inductor on the diode, that diode can’t turn off quickly, so it generates little noise.)
Random thought thread I suppose. Sorry if it disrupts.
This could give you massive power reserve with essentially dual mono topology. Caps would be dramatically smaller for less ripple, and that ripple frequency would be way above audio band.
I used to design switchers, back in the day when 100 kHz was the norm (these days it’s much higher). We were very concerned with noise, as these switchers drove codec linecards for telephony (idle channel noise a big yield problem). So we regulated the secondary of the switcher with magnetic amplifiers (small high permeability toroids)...which have the added benefit of forcing all rectifiers into soft reverse recovery...almost zero noise.
Eventually, I discovered that we could make the noise irrelevant by synchronizing the switching freq to the codec framing pulse, but we had already managed to noise to low levels enough with magamps.
(A mag amp is a small toroid directly in output rectifier path with a control winding on it, such that during the off time of the rectifier, you can preset the remanence point of the magnetic. This allows you pulse width control at the next on cycle. Because it’s a series inductor on the diode, that diode can’t turn off quickly, so it generates little noise.)
Random thought thread I suppose. Sorry if it disrupts.
If you have the Quasimodo ( or cheapmodo) I’d still test each transfomer. They get bought in batches by amp manufacturers and maybe from different suppliers for different batches so snubbing values may change.
Chris, boards and parts are still being revised by Bob and Rick.
I’m looking forward to them finishing so I can renew several Hafler ( and hopefully an old Perreaux with similar Hitachi TO3 Mosfet output stage). Once Bob’s update gets released the price and availability of Haflers will become competitive, best to get one soon.... some old Ashley and Soundcraftsman too.
Your welcome to test or build . I bought parts to rebuild 7 of them, so plenty of semiconductors around for matching.....
Chris, boards and parts are still being revised by Bob and Rick.
I’m looking forward to them finishing so I can renew several Hafler ( and hopefully an old Perreaux with similar Hitachi TO3 Mosfet output stage). Once Bob’s update gets released the price and availability of Haflers will become competitive, best to get one soon.... some old Ashley and Soundcraftsman too.
Your welcome to test or build . I bought parts to rebuild 7 of them, so plenty of semiconductors around for matching.....
Last edited:
Two of my DH-500's have NTC thermistors, but I only know their Hafler P/N, which doesn't correspond to anything I can find. I will be using the DIY softstart board with my new mods--I prefer to completely remove the limiting component from the circuit once the caps are charged. As far as I know only the later model DH-500's came stock with a thermistor. But if adding larger caps to any of these amps it is the least that should be done.
As far as binding posts, Parts Express has a good selection, Binding Posts - Parts Express . I do not have a particular recommendation.
Ron
Speaker Outputs
I do not like binding post at all. I suggest using Neutral SpeakOn chassis mount connectors and the corresponding cable end connectors; or screw terminal blocks. Cramolin the blocks to clean them, and see if you can find plated copper replacement screws.
I do not like binding post at all. I suggest using Neutral SpeakOn chassis mount connectors and the corresponding cable end connectors; or screw terminal blocks. Cramolin the blocks to clean them, and see if you can find plated copper replacement screws.
Yes, it is always a good idea to short the NTC with a relay once power good is detected on the secondary voltages. This removes the remainder of the resistance from the line for in-service operation.
There are two related potential benefits. The NTC relies on being hot to get its resistance down after turn-on. I have not measured this, but it would seem that some amplifiers that do not take a lot of current in the idle state, or when playing very softly, may not pull enough line current to get the NTC at its full operating temperature to get its resistance down to the minimum, so when a sudden loud passage comes along the line resistance due to the NTC may be higher than necessary, depriving the amplifier momentarily of supplying the maximum VA that it would be capable of without the NTC. Always bear in mind that the NTC depends on there being a voltage drop across it to keep it hot and have low resistance.
Also, if the amplifier has been running and the NTC hot, if the amplifier is turned off and then back on quickly, the NTC may still be hot and will not do its full inrush current limiting job because it may not have started the turn-on process when cold.
These two phenomena depend somewhat on the temperature time constant of the NTC, and I have not looked into that time constant.
Cheers,
Bob
Hi Peter,
Any differential pair should be thermally coupled to it's mate.
Close beta matches on the input diff pair will reduce the amount of distortion by a large amount. It tends to sound better too (not surprising). There are ways to destroy the performance of an amplifier so that matching doesn't matter. Transistors closely matched will also need to use matched emitter degeneration resistors as they will throw the match off as well. I was initially surprised by this until I realized that I had matches closer than 1%, then it made sense.
Several people advocate the matched E-B voltage drop method. My own tests do not support matching this way. However, beta matched transistors from the same batch tend to have very similar voltage drops. I guess that E-B matching has a weak correlation whereas beta matching has a very strong correlation for reducing distortion (and also DC offset).
For matching power transistors, I made a large heat sink that you can mount four devices on at a time and use the same principles, or for the same base current, look for similar emitter currents (or collector currents). The critical thing that matching signal or power transistors share is keeping the two transistor die at the same temperature. That or as close as you reasonably can. This is the reason why using the Beta measure function of your meter doesn't match transistors very well. It is critical that you maintain the transistors at the same temperature. The best I can do is to keep the case temperatures very similar. This is all you can do in practice anyway.
-Chris
-Chris
Several years ago I gave a matching jig design to everyone at DIYAudio. I had used it for at least a decade previous to that. What it does is place the pair of transistors in a long tailed pair configuration. They are thermally coupled and isolated from temperature changes outside. Once they settle, all you need to do is measure the voltage between the collectors to determine how well they are matched. Therefore you can use a cheap meter as long as it's zero is accurate because you are using the indicator (meter) as a null detection device. The voltage differentials closer to zero are the closer matches. The jig uses a constant current source that is adjustable to something close to your target current in circuit. At least two members have designed PCBs for it. The information is near the end of one of the Adcom 565 threads, or if you PM me with your email address, I can send you the information I have on it.
Any differential pair should be thermally coupled to it's mate.
Close beta matches on the input diff pair will reduce the amount of distortion by a large amount. It tends to sound better too (not surprising). There are ways to destroy the performance of an amplifier so that matching doesn't matter. Transistors closely matched will also need to use matched emitter degeneration resistors as they will throw the match off as well. I was initially surprised by this until I realized that I had matches closer than 1%, then it made sense.
Several people advocate the matched E-B voltage drop method. My own tests do not support matching this way. However, beta matched transistors from the same batch tend to have very similar voltage drops. I guess that E-B matching has a weak correlation whereas beta matching has a very strong correlation for reducing distortion (and also DC offset).
For matching power transistors, I made a large heat sink that you can mount four devices on at a time and use the same principles, or for the same base current, look for similar emitter currents (or collector currents). The critical thing that matching signal or power transistors share is keeping the two transistor die at the same temperature. That or as close as you reasonably can. This is the reason why using the Beta measure function of your meter doesn't match transistors very well. It is critical that you maintain the transistors at the same temperature. The best I can do is to keep the case temperatures very similar. This is all you can do in practice anyway.
-Chris
-Chris
A voltage-driven differential pair driving a current mirror load is dependent mainly on the accurate balance of the current mirror for its performance - its probably better to match the current mirror transistors and degeneration resistors than those of the pair itself. Of course matching the pair as well will reduce offset voltage, but doesn't contribute to reducing distortion much, as the equal currents do that.
Hi Mark,
Nope. An unmatched current mirror can only degrade the performance of a matched pair. It cannot fix what an unbalanced diff pair does.
The differential pair is the heart and soul of an amplifier. It is here where the input signal is compared to the scaled down (normally) version of the output. This is where the distortion products are pulled out and sent inverted to correct the non-linearity of the amplifier circuit in it's entirety. This absolutely depends upon the two devices, be they tubes or transistors, be matched as closely as possible. By extension this also means the two devices in the diff pair are held at the same temperature. This is also why the higher the transconductance of that stage, the more feedback is available to correct problems (thank you to Walt Jung in private correspondence).
If the diff pair match is corrupted by any means (poor match, dissimilar temperatures, unmatched degeneration resistors or current mirror problems), the distortion products will be incompletely cancelled and the amplifier circuit will perform more poorly than it otherwise could. By the same token, the more linear the transfer characteristic of the rest of the amplifier, the lower the distortion will be.
Perhaps one of the greatest areas that will generate distortion will be the output stage. So matching the output transistors will also lower the distortion as there is less work for the diff pair to do in terms of cancellation. This is one of the strengths of a CFP output stage, and a weakness of a Stasis type design.
Best, Chris
Nope. An unmatched current mirror can only degrade the performance of a matched pair. It cannot fix what an unbalanced diff pair does.
The differential pair is the heart and soul of an amplifier. It is here where the input signal is compared to the scaled down (normally) version of the output. This is where the distortion products are pulled out and sent inverted to correct the non-linearity of the amplifier circuit in it's entirety. This absolutely depends upon the two devices, be they tubes or transistors, be matched as closely as possible. By extension this also means the two devices in the diff pair are held at the same temperature. This is also why the higher the transconductance of that stage, the more feedback is available to correct problems (thank you to Walt Jung in private correspondence).
If the diff pair match is corrupted by any means (poor match, dissimilar temperatures, unmatched degeneration resistors or current mirror problems), the distortion products will be incompletely cancelled and the amplifier circuit will perform more poorly than it otherwise could. By the same token, the more linear the transfer characteristic of the rest of the amplifier, the lower the distortion will be.
Perhaps one of the greatest areas that will generate distortion will be the output stage. So matching the output transistors will also lower the distortion as there is less work for the diff pair to do in terms of cancellation. This is one of the strengths of a CFP output stage, and a weakness of a Stasis type design.
Best, Chris
Thanks for this, I did do coarse matching on the L100 I am rebuilding, think I will re-do that with a more elaborate match as you suggest.
Merry Xmas all!
Hi Peter,
Matching transistors is a pain in the rear, but it is more than worthwhile. For a well matched pair, you should see the DC offset sitting at whatever the design dictates (some designs can have as much as 250 mV that I have seen), hopefully 0 mV or close to it.
Practically speaking, matching the two diff pairs should result in a > 5mV DC offset with this design. If your offset is higher, either your diff pairs are not matched, or you have other problems with that module.
-Chris
Matching transistors is a pain in the rear, but it is more than worthwhile. For a well matched pair, you should see the DC offset sitting at whatever the design dictates (some designs can have as much as 250 mV that I have seen), hopefully 0 mV or close to it.
Practically speaking, matching the two diff pairs should result in a > 5mV DC offset with this design. If your offset is higher, either your diff pairs are not matched, or you have other problems with that module.
-Chris
Chris, Chris ...
Still on to the merits of matching devices vs performance. I which you could come up with some tests/measurements &/or examples to demonstrate your claims. In another thread please
How well can you match a pair of MJL1301A & MJL3281A devices for example? With semi processes being so well controlled these days finding hfe matches between npn & pnp devices is near impossible even if it is a requirement you will not be able to meet your own design requirements with current production of parts.
Since this is a "Hafler DH-200/220 Mods" thread, how do your claims to matching devices apply to the DH-200/220 design?
Requirement and how for example for matching the 2sk134 & 2sj49 devices and to what specs?
FYI, I asked Bob Cordell about matching output devices. He said that is why he prefers a 3EF design, because the accumulated gain of the 3 EF makes matching unnecessary, since current the gain is in >=500k range.
Happy New Year
Rick
Last edited:
Hi Rick,
Well, Bob and I sharply disagree on this then with regard to matching input devices. I have proved this to be effective continuously over 30 + years. So it doesn't bother me at all that anyone might disagree with me as I have years of accumulated proof. I have measured significant drops in distortion simply by installing a close matched pair into an amplifier or gain stage.
Matching outputs helps if you can make the overall gain in an output double or triple to come close to each other. The reasons for this should be obvious. The less correction that needs to be made to the signal, the lower the overall distortion will be. This is because any gain stage has only a certain amount of excess gain that can be used to correct errors. This falls as the frequency is raised (no surprise to anyone I trust).
Matching something like NPN outputs, and then the group of PNP devices becomes important when you have two or more output pairs per channel. Having the groups match is icing on the cake if you can get that to happen.
In the Hafler DH-200 / 220 design, matching the complimentary input pairs brings rather large benefits. This is another thing I have measured over the years. The side benefit is low DC offset, which isn't strictly that important below some arbitrary number ~ say |50 mV| for example. You also want the tail currents to roughly match as well.
When I rebuild one of these amplifiers, I match the input pairs and select a couple of complimentary pairs as reasonably close as I can. Then the tail current sources have roughly matched transistors. Then I match the TO-5 devices, again as reasonably close as I can. The outputs are left as they are, they are low transconductance devices anyway.
When I buy transistors, they are purchased in lots of 100 normally. The TO-5 parts are bought in groups of 20 or 30 depending on how rich I feel at that moment. I do enough work so that these quantities are affordable because by matching, I do a better repair. People have commented on the improvement in sound quality without any of this being mentioned. DH-101 preamps also greatly benefit from this treatment.
I tend not to use MJL1302 type outputs, I do use MJL21195 types. The same polarity matches are easy to get, especially compared to what I had to deal with from the 70's on through to the introduction of the newer On-Semi parts. Back then it was all about 2SD424 and 2SD555 parts that I stocked as they were more than good enough for most amplifiers. I think the Marantz 300DC use 2SD665 or something that would be ordered in bunches when I needed them. I also used MJ15024 types, but I had to order a lot of those to get matches. For special semis, I ordered them in matches sets from the distributor when required.
When you don't bother with device matching, you may get current hogging and higher distortion. This higher distortion is only reduced by the amount of free gain above what is needed for the feedback loop. So the distortion is still there, but could have been reduced to lower levels (~20 dB according to On-Semi) by matching the output transistors. Hey! They wrote that, I didn't. That 20 dB figure is before feedback is applied, so some designs will have higher residual distortion than others for the same amount of feedback. This would be the amount of excess gain that can be used to reduce distortion. Generally, the less sophisticated designs will benefit more from matching than sophisticated ones. But, don't those better designs deserve to operate with the least amount of distortion possible? Isn't that why they were purchased in the first place? The Hafler amplifiers are not sophisticated, that's how all this ties into the topic of the thread. That and I am responding to your comments.
With regard to input matching, I believe that Douglas Self agrees with me. I'll have to read through Bob's book to see if he also agrees with this or not. As for matching output transistors, this is much easier these days than it ever has been before. So manufacture-ability is enhanced. Looking at it another way, it is more difficult to end up with severe current hogging today than it was. In the past, power devices were graded at the factory for really rough matches, that's how bad it was. But look in many service manuals and you will see references to using graded parts, or confirming that bias current is equally shared among output devices (= matched parts). The evidence is all around you if you care to look. I guess that if you aren't a technician (a good one), this is news to you, but that doesn't mean this isn't an issue from way back. The Marantz 500 amplifier, sold from 1969 to 1974, specified in the manual that the outputs and many other parts must be matched. Download it from anywhere and look at the last pages. That means they knew about this from at least 1968 and forward. Why 1968? Because the 500 was designed at least one year prior to it's manufacture and the publication of the service manual.
I didn't grab these ideas out from thin air. In my life as an audio technician, I have been surrounded by this information and taught the same thing by several different companies and technicians. They all differ in presentation, but the actual kernel of what was taught is that output transistors must be matched for top performance, and even for the simple reason that they might fail if this is not done - depending on how bad the mis-match is. Perhaps the most detailed and informative education I received from a manufacturer was Marantz (Superscope at the time). They also showed me that not only is it better to match the diff pair, but even the complimentary transistors through the design from input to output. If you think about it, it even makes sense.
Sorry for the long response, but there is good information in there. Your job, use it.
-Chris
Well, Bob and I sharply disagree on this then with regard to matching input devices. I have proved this to be effective continuously over 30 + years. So it doesn't bother me at all that anyone might disagree with me as I have years of accumulated proof. I have measured significant drops in distortion simply by installing a close matched pair into an amplifier or gain stage.
Matching outputs helps if you can make the overall gain in an output double or triple to come close to each other. The reasons for this should be obvious. The less correction that needs to be made to the signal, the lower the overall distortion will be. This is because any gain stage has only a certain amount of excess gain that can be used to correct errors. This falls as the frequency is raised (no surprise to anyone I trust).
Matching something like NPN outputs, and then the group of PNP devices becomes important when you have two or more output pairs per channel. Having the groups match is icing on the cake if you can get that to happen.
In the Hafler DH-200 / 220 design, matching the complimentary input pairs brings rather large benefits. This is another thing I have measured over the years. The side benefit is low DC offset, which isn't strictly that important below some arbitrary number ~ say |50 mV| for example. You also want the tail currents to roughly match as well.
When I rebuild one of these amplifiers, I match the input pairs and select a couple of complimentary pairs as reasonably close as I can. Then the tail current sources have roughly matched transistors. Then I match the TO-5 devices, again as reasonably close as I can. The outputs are left as they are, they are low transconductance devices anyway.
When I buy transistors, they are purchased in lots of 100 normally. The TO-5 parts are bought in groups of 20 or 30 depending on how rich I feel at that moment. I do enough work so that these quantities are affordable because by matching, I do a better repair. People have commented on the improvement in sound quality without any of this being mentioned. DH-101 preamps also greatly benefit from this treatment.
I tend not to use MJL1302 type outputs, I do use MJL21195 types. The same polarity matches are easy to get, especially compared to what I had to deal with from the 70's on through to the introduction of the newer On-Semi parts. Back then it was all about 2SD424 and 2SD555 parts that I stocked as they were more than good enough for most amplifiers. I think the Marantz 300DC use 2SD665 or something that would be ordered in bunches when I needed them. I also used MJ15024 types, but I had to order a lot of those to get matches. For special semis, I ordered them in matches sets from the distributor when required.
When you don't bother with device matching, you may get current hogging and higher distortion. This higher distortion is only reduced by the amount of free gain above what is needed for the feedback loop. So the distortion is still there, but could have been reduced to lower levels (~20 dB according to On-Semi) by matching the output transistors. Hey! They wrote that, I didn't. That 20 dB figure is before feedback is applied, so some designs will have higher residual distortion than others for the same amount of feedback. This would be the amount of excess gain that can be used to reduce distortion. Generally, the less sophisticated designs will benefit more from matching than sophisticated ones. But, don't those better designs deserve to operate with the least amount of distortion possible? Isn't that why they were purchased in the first place? The Hafler amplifiers are not sophisticated, that's how all this ties into the topic of the thread. That and I am responding to your comments.
With regard to input matching, I believe that Douglas Self agrees with me. I'll have to read through Bob's book to see if he also agrees with this or not. As for matching output transistors, this is much easier these days than it ever has been before. So manufacture-ability is enhanced. Looking at it another way, it is more difficult to end up with severe current hogging today than it was. In the past, power devices were graded at the factory for really rough matches, that's how bad it was. But look in many service manuals and you will see references to using graded parts, or confirming that bias current is equally shared among output devices (= matched parts). The evidence is all around you if you care to look. I guess that if you aren't a technician (a good one), this is news to you, but that doesn't mean this isn't an issue from way back. The Marantz 500 amplifier, sold from 1969 to 1974, specified in the manual that the outputs and many other parts must be matched. Download it from anywhere and look at the last pages. That means they knew about this from at least 1968 and forward. Why 1968? Because the 500 was designed at least one year prior to it's manufacture and the publication of the service manual.
I didn't grab these ideas out from thin air. In my life as an audio technician, I have been surrounded by this information and taught the same thing by several different companies and technicians. They all differ in presentation, but the actual kernel of what was taught is that output transistors must be matched for top performance, and even for the simple reason that they might fail if this is not done - depending on how bad the mis-match is. Perhaps the most detailed and informative education I received from a manufacturer was Marantz (Superscope at the time). They also showed me that not only is it better to match the diff pair, but even the complimentary transistors through the design from input to output. If you think about it, it even makes sense.
Sorry for the long response, but there is good information in there. Your job, use it.
-Chris
I should have known better than to open up this can of worms with you
lets get back to DH-200/220 mods please,
I have the DH-220C-A1 fired up and running on the bench. I put in all Vbe matched devices except the KSA1381E and KSC3503D, they do not come in the same hfe grading anymore, chances of find a match are like winning the lottery = waste of time & $.
You guys are not reading what I asked about matching npn & pnp (n-ch & p-ch) not like polarities, that I thought was obvious.
lets get back to DH-200/220 mods please,
I have the DH-220C-A1 fired up and running on the bench. I put in all Vbe matched devices except the KSA1381E and KSC3503D, they do not come in the same hfe grading anymore, chances of find a match are like winning the lottery = waste of time & $.
You guys are not reading what I asked about matching npn & pnp (n-ch & p-ch) not like polarities, that I thought was obvious.
Attachments
Last edited: