I have built this kind of amplifiers , but I need to match the LTP transistors and i cant measure their hfe ( my multimeter's socket it's broken ) , and also I dont have an osciloscope.
I have built Rod elliot's P3A , didnt worked well because of the LTP missmatch
I have built Rod elliot's P3A , didnt worked well because of the LTP missmatch
Nope. Mismatch doesn't matter. Use the same resistance to ground for each base of the LTP, and use a current source in the tail, and you're good to go. The P3A is decent, but I prefer a current source in the VAS. I've also built similar amplifiers to the P3A, and they also don't work as well as my other designs. With his P3A, PCB layout is important, and for any amplifier, grounding is very important (as well as earthing, if you're earthing).
My main designs are quite similar to the P3A in many ways, but I have a number of simple differences which make my noise performance incredibly good. I believe in a good CMRR in the LTP - it's the only way to get the best noise performance. But using a current mirror leads to more instability. So use clever resistance values with a clever zener in the LTP current source, and you end up with excellent CMRR. Current source in the VAS and well calculated current values, and the noise performance holds. My final design is now at 103 dB SNR and about 0.005% THD. The design is independent of Hfe (within a certain tolerance, obviously).
This preamplifier has the same principals. Simulate it - use a 3055 in one side of the LTP, and a BC on the other.
i dont know then 😀, but at closest o 0 mv dc offset it had distortion, if the dc offset was higher than 500mV ( LOL ) it sounded better ..
I have built the schematic in post 1 , sounded good , no oscilations ( i think, dont have osciloscope to check, but didnt hear distortions / oscilations on the speaker , not on 10-20 hz or 10khz + . But the maximum power was , 8-9 v on music , and 12v on 1khz down to 20hz .
DC offset shouldn't make that much difference to distortion.
Back to op amps - the TIP147s were probably the cause of lower output at higher frequencies. I'd say a little class A op amp amplifier might not be a bad idea.
Back to op amps - the TIP147s were probably the cause of lower output at higher frequencies. I'd say a little class A op amp amplifier might not be a bad idea.
Basically the same design can be found in the 1975 RCA manuals, called Level Translator. TR7 and TR 16 act as common base amplifiers, with their bases about 3V apart. They're driven by their emitters through R9/C13. TR2 and TR10 are common emitter drivers and voltage amplifiers. This complete section, including the power devices, provides a voltage gain of R45/R9 = 3,7. As this is an inverting amplifier, the OpAmp's non-inverting input has to be used for NFB.
Ah, an Elektor design from the 1990ies! Yes, as T1's and T3's bases virtually are tied together, only one compensation cap is needed here. T4 and T%, with their surrounding components, are the usual VI limiters.
Best regards!
You don't need a scope.
And the DMM hFE is only good for rough batching of device hFE.
You really should be using a jig that allows you to set the correct operating current (Ie) to match your amplifier's operating current. Then you measure the base current that got you to the required Ie. That gives hFE at your operating current.
For an LTP requiring matched pairs you measure the Vbe when the TWO devices are in a dual jig (thermally coupled) with zero ohms base stopper and passing nearly the same Ie through the two collector loads. It's the nearly equal Ic that now tells you they have nearly the same Vbe. If Ic were identical, then you have identical Vbe at that current.
It this this second part of matching that I and most Members of the Forum do not agree.
They claim that hFE matching is sufficient and that Vbe then ends up being good enough.
But if your amplifier, as pointed out by Mrcloc, does not benefit from matching then you don't need to bother.
Some LTP pairs are deliberately mismatched to increase the proportion of 2nd harmonic and this also increases the other even harmonics slightly. This is claimed by some Designers to give a "nice sounding" amplifier preferred by many listeners.
Last edited:
I agree with this, and there is merit to doing Hfe matching in certain cases, but as a general rule, if your amplifier needs Hfe (or Vbe) matched devices to be reliable, the amplifier is unreliable.
Take the basic preamplifier circuit I posted. Add a second NPN to one of the LTP NPNs like this:
Now Hfe and Vbe are very different. The key to an LTP is to balance the current in each collector. The closer the Ic, the better the CMRR. With that circuit above, Ic for each side remains very similar.
Hfe isn't in any of my equations; there just has to be enough, and remain as constant as possible within the Q point of each transistor.
Take the basic preamplifier circuit I posted. Add a second NPN to one of the LTP NPNs like this:
An externally hosted image should be here but it was not working when we last tested it.
Now Hfe and Vbe are very different. The key to an LTP is to balance the current in each collector. The closer the Ic, the better the CMRR. With that circuit above, Ic for each side remains very similar.
An externally hosted image should be here but it was not working when we last tested it.
Hfe isn't in any of my equations; there just has to be enough, and remain as constant as possible within the Q point of each transistor.
On the circuit in post 67, T4 & T6 and D5 & D6 are the VI limter. this is an obsolete 80's kind of protection device for the problem of 1/4 phone plugs to the speaker pulling part way out and shorting out the output transistors. Modern amps use speakon connectors that don't pull out. Also the modrn solution to shorted output is current detect & speaker disconnect with MOSFET's (see protection circuit threads) In the Marshall 8008 TR6 & TR15 perform the same VI limiter function.
VI limiters measure the current across the emitter resistors of the output transistors, and if it gets over .6 v, the limiter transistor turns on and diverts the base drive current to the driver transistors, to ground. So the output transistors aren't driven so hard.
The Marshall 8008 IMHO drives the VAS with +-15 because it allows them to use cheap VAS transistors with low Vce. lower VAS supply also limits clipping power too, which guitar amp users like to do but I never do. Clipping in the output stage can melt the speaker coil, & that causes warrenty returns. See Harley Peavey white paper about his speaker coil warrenty returns in the early days causing his shop to invent the DDT circuit. The Marshall limited VAS supply voltage is cheaper since they needed +-15 for the op amp anyway.
VI limiters measure the current across the emitter resistors of the output transistors, and if it gets over .6 v, the limiter transistor turns on and diverts the base drive current to the driver transistors, to ground. So the output transistors aren't driven so hard.
The Marshall 8008 IMHO drives the VAS with +-15 because it allows them to use cheap VAS transistors with low Vce. lower VAS supply also limits clipping power too, which guitar amp users like to do but I never do. Clipping in the output stage can melt the speaker coil, & that causes warrenty returns. See Harley Peavey white paper about his speaker coil warrenty returns in the early days causing his shop to invent the DDT circuit. The Marshall limited VAS supply voltage is cheaper since they needed +-15 for the op amp anyway.
Last edited:
- Status
- Not open for further replies.