hitsware said:
No, just a simple opamp with PNP LTP, NPN LTP, and VAS stages loaded on each other.
I took one and connected inverting input to output for an unity gain.
Then, I used resistors in tails, so max output current would depend on a voltage swing. Then, I replaced inverting output transistors by diodes, because their b-e junctions break down on too low voltages for my application. Since an output resistance is low, transistors on inverting inputs are not necessary, so diodes with higher break-down voltages may be used.
Such a way I made a class A opamp that when load resistance decreases it gradually turns from a voltage source to a current source. Then, for a power efficiency I added diodes with resistors in series in tails, turning it into class AB amp.
This is an example. Earlier, I was wondering why Claude had to use diodes where I used transistors. Then I recalled the case when I myself used diodes instead of transistors.
Sharing such examples we are learning.
I knew about exceptional Vishay bulk foil resistors, but I did not know about their exceptional diodes. Now thanks to Claude I know them so may use when needed.
hitsware said:
I had then some selected KT-818 and KT-819 transistors I obtained earlier in Tomsk Institute of Semiconductor Devices. I selected them from a batch with 60V Vcbe, but all of them had H21=200 and Vcbo>150, if I demember right now (time passes!). Also, I had a 150W toroid I bought on a flea market to build a HI-Fi for home. Having no time to rewind a guitar amp's output transformer I made a quick fix from what I had. As I know, the owner did not come back to EL-34 output tubes, he used my quick fix ears later.
How to adjust.
1. Turn all pots on Max resistance.
2. Connect a dummy load corresponding to the maximal current needed.
3. Connect an oscillator and bring down both class B pots until you get no clipping on max voltage.
4. Brigg down both pots for class A bias (long tails to rails) until crossover distortions disappear.
Turning pots watch for symmetry (integrating R-C network with DC measurement tool connected).
However, it required careful adjustments, but I could not dream to rewind a burnt output transformer overnight.
Wavebourn,
I`m somewhat less enthusiastic about diodes, there`s no reason to glorify them, they are anything but linear (including 1N914B), doing a great job when various nonlinear effects are desired. This is not a big problem in a guitar amp, where linearity is rather a nuisance. The best diode is no diode. Diodes are to a large extent avoidable with a purposeful attitude.
I`m somewhat less enthusiastic about diodes, there`s no reason to glorify them, they are anything but linear (including 1N914B), doing a great job when various nonlinear effects are desired. This is not a big problem in a guitar amp, where linearity is rather a nuisance. The best diode is no diode. Diodes are to a large extent avoidable with a purposeful attitude.
Lumba, you are right. Diodes are to be avoided, if possible, as they tend to create higher order non-linearities. However, as a design exercise, for an engineer, it can be interesting, at least for a time. Sprinkling diodes in a discrete design seems pointless, although in IC's, I can easily see where they could promote linearity, if used properly.
Lumba Ogir said:
I am totally agree with you. Diodes are devices with exponential curves. But what are transistors?
The same.
You may view an emitter follower as a diode with feedback by voltage. As the result, the diode works on the shorter part of it's characteristic. But anyway it is a diode.
Base - emitter junction is a diode with the same exponential curve. Driving base-emitter junction you drive a diode.
Collector-base junction is again a reverse biased diode that exhibits current proportional to temperature.
In usual common emitter connection driving base-emitter diode you change properties of base - collector diode...
Changing voltage base-collector of that reverse biased diode you also change properties of base - emitter diode... I.e. amplifying signal it changes own input properties!
Also, as any diodes they have varactor effect!
Transistors have to be avoided!
Like bees... They bite.
Actually, Wavebourn, that is an unproductive statement. Yes, transistors have diode like characteristics, BUT they can be linearized by using a diode at the input, as shown convincingly by Barrie Gilbert, decades ago. Also, transistors can be 'linearized' by adding an emitter resistor and use it to heavily modify the input behavior. You can still get gain, while you can't with just a diode.
However, putting in diodes instead of resistors, or transistors in most spaces makes little sense.
However, putting in diodes instead of resistors, or transistors in most spaces makes little sense.
John,
A little earlier you mentioned some very special Hitachi bipolars but couldn't remeber the numbers.
The ones we had recommended to us back around 1975 as being VERY special were the Hitachi 2SA1085 PNP and 2SC2547 NPN.
They were astonishingly linear and sounded great as straight replacements for regular Motorola Trs. We still use them in our solid state SACD/RBCD player Upgrade modules.
Regards, Allen (Vacuum State)
A little earlier you mentioned some very special Hitachi bipolars but couldn't remeber the numbers.
The ones we had recommended to us back around 1975 as being VERY special were the Hitachi 2SA1085 PNP and 2SC2547 NPN.
They were astonishingly linear and sounded great as straight replacements for regular Motorola Trs. We still use them in our solid state SACD/RBCD player Upgrade modules.
Regards, Allen (Vacuum State)
That could well be them. I picked up the numbers in 1978, from a technical magazine that I picked up in Japan. These darn things are amazing! However, they have a 'kink' in their transfer function, unless you operate them with at least a 1V or 2. You can see it on a curve tracer. It made it useless in my patented low noise input topology that I was using at the time.
Sanyo does almost perfect equivalents under 2SC2362/2SA1016 (if memory serves). They are indeed very quiet, I measured an rbb of under 2ohm. I'm using them to experiment with a low noise MC input stage. Got them from www.bdent.com they cost pennies.
I found a fair number of devices with low Rbb' by 1980, but many have high input capacitance and medium to low beta. Not so, with the Hitachi devices.
Again, and Syn08, this is something that you might explain, why did these Hitachi devices fail to perform in regards to distortion when operating with 0 Vcb? Any ideas? I keep thinking, extra diffusion, but this is not my territory.
Again, and Syn08, this is something that you might explain, why did these Hitachi devices fail to perform in regards to distortion when operating with 0 Vcb? Any ideas? I keep thinking, extra diffusion, but this is not my territory.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.