2 Terry
Put in series with colector of Q7 resistor about 15k, it will shift power dissipation on this resistor..Increased current in LTP is needed to improve linearity. The same is it for VAS, it is impossible to run VAS at current 1,5-2mA, linearity will be than very poor, minimum is about 5-10mA. Transistors with TO-126 package is a must as minimum for VAS with Ub+-90V, here will be dissipation at least 0,5-0,6W and more.
In your original schematic, VAS loaded with resistor was working in classB...
Put in series with colector of Q7 resistor about 15k, it will shift power dissipation on this resistor..Increased current in LTP is needed to improve linearity. The same is it for VAS, it is impossible to run VAS at current 1,5-2mA, linearity will be than very poor, minimum is about 5-10mA. Transistors with TO-126 package is a must as minimum for VAS with Ub+-90V, here will be dissipation at least 0,5-0,6W and more.
In your original schematic, VAS loaded with resistor was working in classB...
Last edited:
You've got some real issues going on there that need figuring out.
The wiring up of amplifiers, with respect to return currents, both in terms of wiring up the amplifier for every day use and then wiring it up for measurements, is critical.
The results that you've posted clearly show that there is something contaminating the results from how they should otherwise be.
As a simple test could you perform a measurement, like above, 14VRMS out, but without any load connected?
I do what i can ,if it's wrong i will try again.No wonder you've gave us so controversial reports.
I think better way is if you give me some advice.
I'm not a professor.
Last edited:
OK I've had some time to look at BV's spice model. As well as having 3.4mV across Q7 I see 10mv across R10 which I believe is the VAS output. Am I wrong here or is the VAS output half that? I can see that this Symasym board is going to end up being a conventional Symasym because it just is not adequate for 70V rails. Still it is working as a learning tool which is great.
Blessings, Terry
Blessings, Terry
This may seem like an odd question having just asked about the PSRR of the VAS of the CXA-H, but really how good/bad is it? I am currently laying out a board and as these are going to be run from lowish voltage rails I am going to fully regulate the supply for the input stage.
The question is will the VAS really benefit from a similar approach? I mean you could say that any VAS, regardless of its PSRR, will benefit, but is it likely to actually improve the measured performance of the amp? Bearing in mind that these are going to operate in class A, which should do away with half sine pulses from the power rails of the output stage.
The question is will the VAS really benefit from a similar approach? I mean you could say that any VAS, regardless of its PSRR, will benefit, but is it likely to actually improve the measured performance of the amp? Bearing in mind that these are going to operate in class A, which should do away with half sine pulses from the power rails of the output stage.
Without the cap multiplier , most the CFA's are (-) 60-70db across the audio
range while the VFA's are near - 100db.
This is inherent to the different topologies ... nothing can be really done
about it except to filter the rails , regulate and current source the input
pair , or "float" the whole deal on current sources.
The multiplier gives another 10 - 12db PSSR to give around 80db generally.
Since the multiplier is common to all my IPS's, some of the VFA's will
even simulate well over -120db (wolverine 125+).
-80db is on par with a bootstrapped LIN amp ... not TOO bad.
PS - this is why the "slewmaster OPS" has the option for a separate/boosted
supply (those jumpers near the driver stage) - to allow for a fully regulated
supply from the main drivers back to the IPS.
Terry ,yes- the VAS output current will be half what the main Re current is.
Half is used by the "backend" (level shifter).
OS
OK as usual I need some help.
I couldn't find a BOM anywhere for the CFAX-H so I used the schematic. I actually put together a spice file so I could try to determine the voltages on the caps but I couldn't get it to run. I want to test the board before I hook it up to the OPS but I can't seem to find any details about how to do the initial setup. I'll attach the spice file so maybe one of you can find my error so I can have a basis to compare to.
Thanks, Terry
I couldn't find a BOM anywhere for the CFAX-H so I used the schematic. I actually put together a spice file so I could try to determine the voltages on the caps but I couldn't get it to run. I want to test the board before I hook it up to the OPS but I can't seem to find any details about how to do the initial setup. I'll attach the spice file so maybe one of you can find my error so I can have a basis to compare to.
Thanks, Terry
Last edited:
It is inherent to usual solutions, but it is possible to make some "subtle" changes in topology (no need for multiplier in rails , simple RC filtering for input stage is more than enough) and so improve PSRR more than 40dB.
The only proof about real PSSR is measurement, at various power levels without visible (always below 120dB ref. signal) mains artefacts
You should just need to centre the pots and adjust one or the other to null the offset. By design the currents should be close to the design target. That worked fine for me anyhow.
Member
Joined 2009
Paid Member
a) Mosfets although does not provide DC load at all, but for AC (audio signal) here is very nonlinear input capacitance to charge and discharge, depending on speed and output load. For higher frequencies and paralleled MOSFETs , needed currents are almost the same as for bipolars.. So reduce VAS load , isolate VAS from nonlinear load , is allways advantageous for VAS linearity.
For most stringent requirements MOSFET needs some kind of error correction , and such circuit always provide aditional load for VAS.
b) sound depends (IMO..) on measured performance, so no comments more here
Last edited: