Thanks Valery,
I will try that this morning. Hopefully that will make this pair usable. I think I will build a pair with MJL4381/4302 and see how those work out. I asked about those before building this pair and was told the MJL21193/94 might be better at the higher voltage because of their higher current capabilities. I will try your cap suggestions and report back. The Sankens would probably work but on these boards you would have to drop down to three pair in order for them to fit.
Blessings, Terry
I will try that this morning. Hopefully that will make this pair usable. I think I will build a pair with MJL4381/4302 and see how those work out. I asked about those before building this pair and was told the MJL21193/94 might be better at the higher voltage because of their higher current capabilities. I will try your cap suggestions and report back. The Sankens would probably work but on these boards you would have to drop down to three pair in order for them to fit.
Blessings, Terry
Last edited:
This begs the question then, in a triple EF, do we benefit at all from different device choices for driver and pre-driver, especially to tailor response to the very capable MJL2119X devices? What is optimal with respect to their Ft? Given the choice I'd suggest more compatible drivers and pre-drivers, though slugging them with a capacitor does work. I always figured relative speeds should be:
Pre-Driver - Fastest of the three
Driver - Fast, somewhere between the pre-driver and output speed
Output - Slowest of the three
But perhaps there is a better combination, or we simply want slower drivers and pre-drivers when we are choosing to use slower outputs. Some comments were made some time ago in this thread about it but I'm not sure I can find them now.
You can get Sankens in a to-3P package (IIRC Sanken calls it the MT-100).
I was primarily looking at the SOA when I suggested the MJL2119x devices. I honestly didn't think the speed of the device would become an issue. To be fair, my prototype build did use the faster type of outputs.
For those who want a 3P option the MT-200 Sankens would be really nice...
In fact, I also considered MJL21193/21194 as no problem at a first glance, without even looking at its Ft, and I have them in my SlewMonster prototype (3 pairs as this is the number I could buy quickly) and it does not oscillate in my case. I only started looking deeper after the point was raised here.
But the pole is there, in my case it's just a little bit lower (on vertical axis).
So let's call this a "potential " issue so far.
I did an interesting simulation earlier with regards to a commercial design with CFP topology and enormously heavy compensation - it used rather slow TIP35/36 output transistors and behaved very unstable even with all the 680pF shunt caps in place. Simply changing those to big Sankens allowed running it stable even with no caps at all - same nature, with faster Sankens the pole moves far to the right and down on the graph and this is what we basically want from it...
But the pole is there, in my case it's just a little bit lower (on vertical axis).
So let's call this a "potential " issue so far.
I did an interesting simulation earlier with regards to a commercial design with CFP topology and enormously heavy compensation - it used rather slow TIP35/36 output transistors and behaved very unstable even with all the 680pF shunt caps in place. Simply changing those to big Sankens allowed running it stable even with no caps at all - same nature, with faster Sankens the pole moves far to the right and down on the graph and this is what we basically want from it...
Last edited:
Where we can find a supplier for these MAG....
Mouser haven't them on stock and i read somewhere that Farnell stock some fake 😱
Mouser haven't them on stock and i read somewhere that Farnell stock some fake 😱
Last edited:
Thanks Idefixes ,no one know🙂
I have two pairs 2sc3264,a1295 from the past,i'm sure these are original ,tosted many times at 400Khz and hi volume .
Last edited:
Hi Guys,
I am building a pair of the SlewBaby OPS boards. They are like the Slewmaster OPS except they have only 2 pair outputs and the drivers are TO-220 devices.
My question is should the current be different through the TO-220 devices than the TO-247 devices? Should I use a higher value for R113?
I'm planning to use 2SC3423-Y/A1360-Y for the predrivers and cap mutiplier, BD139 for the VBE multiliers, MJE15032/33 for the drivers and MJL4281/4302 for the outputs. It will most likely be run with +-56v rails. Does anyone see any problems with my device choices? Do you see any other components that need to be changed in the attached schematic?
Thanks, Terry
I am building a pair of the SlewBaby OPS boards. They are like the Slewmaster OPS except they have only 2 pair outputs and the drivers are TO-220 devices.
My question is should the current be different through the TO-220 devices than the TO-247 devices? Should I use a higher value for R113?
I'm planning to use 2SC3423-Y/A1360-Y for the predrivers and cap mutiplier, BD139 for the VBE multiliers, MJE15032/33 for the drivers and MJL4281/4302 for the outputs. It will most likely be run with +-56v rails. Does anyone see any problems with my device choices? Do you see any other components that need to be changed in the attached schematic?
Thanks, Terry
Attachments
Hi Terry, I don't see any problems. No stability issues expected.
Power dissipation of TO-220 drivers will not exceed 1.6W in normal operation (with R113 = 68R quiescent current = 20mA, Pd = 1.12W). It may raise up to 5.5W in case you run a square wave at full power, but a) it is still acceptable assuming a decent heatsink b) it's a very rude way of using an audio amplifier 😉 Anyway, the output devices will already feel much more stressed at that time.
Cheers,
Valery
Power dissipation of TO-220 drivers will not exceed 1.6W in normal operation (with R113 = 68R quiescent current = 20mA, Pd = 1.12W). It may raise up to 5.5W in case you run a square wave at full power, but a) it is still acceptable assuming a decent heatsink b) it's a very rude way of using an audio amplifier 😉 Anyway, the output devices will already feel much more stressed at that time.
Cheers,
Valery
Seems to me that VFAs equally have a pole due to stray capacitance at the -ve input. The lower impedance of this input in a "CFA" should push this pole further out.
So for audio power amps, that are all limited to similar speed by the OPS, this should actually favour the CFA (a little!).
Very different for ICs, as you note, where the speeds of typical "CFA"s make this pole an issue for them, probably more often than for typically slower VFA.
Best wishes
David
Never made serious studies about this, but, at first glance, i would try to make each stage faster than the previous,
to avoid TIM.
At the second look, if the amp is class AB, we would like to have the 'comparator' faster than the output stage to cure crossing distortion ?