Thanks Salas. I'll measure how much current the driver stage draws and what voltages I have to work with so I can work out a value for the resistor in your suggested filter, allowing for the 20V delta for the regulator circuit.
Ray
When you set an SSHV2 for constant current you should allow 20mA extra than the max load demand. For self consumption purposes. Also keep in mind than when touching the test points with the DMM probes they interact keeping back about 5mA from the real setting. Seeing 45mA means that its running at 50mA when untouched for example. Every 10mV TP reading means 1mA current.
As it happens there is a resistor in the power line to the driver stage so I was planning to measure the voltage across it and resolve ohms law to determine the current load of the stage.
Ray
Ray
I meant to type mA indeed. Its certainly not some current that the probes steal. Probably some interaction of their leads inductance and the cascode CCS that tends to show little less TP figure than when not sampled. I had verified that trend using series current measurement in the input and comparing live when measuring on TP with DVM and when not. They both show the same when TP is sampled, but the input current goes a little up when not sampled. Now if its just my measurement gear/set-up, which is a possibility, others may confirm by performing the same experiment (always in LV safe CCS test).
Do it when driven with sinewave signal in good swing so to determine the full RMS current, not just the bias current.
The voltage across the resistor is near constant.
The reactance of a pair of probes cannot affect a DC voltage.
The input impedance/resistance of the meter will affect the value read off the resistor with the DC voltage across it.
It can be modeled by comparing the Resistor value alone to the value of the resistor parallel to the input R. Typically 100k to 10M || 10r gives an error of <0.01%. Most Members checking this voltage would not correct for that less than 0.01% measuring error. It is swamped by other errors.
Probe leads should be twisted from the meter to the probes and just enough free/untwisted to reach across the device being measured. This minimises inductance and maximises capacitance. That capacitance is quite low due the the thick insulation on each of the probe cables.
The reactance of a pair of probes cannot affect a DC voltage.
The input impedance/resistance of the meter will affect the value read off the resistor with the DC voltage across it.
It can be modeled by comparing the Resistor value alone to the value of the resistor parallel to the input R. Typically 100k to 10M || 10r gives an error of <0.01%. Most Members checking this voltage would not correct for that less than 0.01% measuring error. It is swamped by other errors.
Probe leads should be twisted from the meter to the probes and just enough free/untwisted to reach across the device being measured. This minimises inductance and maximises capacitance. That capacitance is quite low due the the thick insulation on each of the probe cables.
Last edited:
The theory is alright, but I have seen the deviation glitch by measurement practice for the given cascode CCS/layout/DMOS type/gate stoppers value. I used Fluke 87V for measuring TP and 1mA confirmed accurate LAB PSU as 20V voltage source/current reader. An artifact or not, its 100% repeatable on my bench. Even with gate protection zeners, diodes, or nothing for D1 D2.
You meant amplitude, not RMS?
It's probably just noise pick-up through EMI which is amplified by the DMOSs.
More likely that there is a stability issue.
The change in parasitic capacitance may be affecting the operation enough to change the output.
If that is the case then the stability margins should be investigated and corrected.
Touching both ends of a low voltage resistor should not affect operation.
The change in parasitic capacitance may be affecting the operation enough to change the output.
If that is the case then the stability margins should be investigated and corrected.
Touching both ends of a low voltage resistor should not affect operation.
Next time I will scope the sampling resistor across to see. Can be the lab psu reacting also. There are long leads involved.
Finished a couple of SSHV for my RTP-3D preamp. I cannot get one of the regs going higher than 280v though the other working fine. I'm targeting 300v/70mA. The voltage after the rectifiers and filtering is about 360v, dummy resitor 5K, 70mA reading on the test points. Can you advise where to look, please?
Either not enough Idss in the Jfet or the load is heavy for the CCS current set. Give 10mA more CCS as first test.
The IDSS is enough for those 68K+68K Vref resistors. Aim for 80mA CCS setting. 5K load at 300V asks for 60mA already. You can drop 40V at 80mA on some CRC from the 360V feed before the reg. So to burn at 20V only on the CCS MOSFET to avoid some heat there. Also see not to have long wires to the reg's input generally. Else, use a film bypass capacitor across the input.
Hi, Salas, I have read a lot of pages of this wonderful thread. Finally I started to build one. I have problem and I wish you can help me.
The SSHV2 I'm building is for my preamp. The preamp runs 290V@5ma per channel. The DC after rectifier is 400V. I used CRC to drop the voltage to around 350V. I used 80K/5W resistor as dummy load for the reg. I understand I should set the R4 to get 0.25V reading on R5 ( 20ma+5ma ) , then adjust R11 to get 290V for output.
My problem is right now, at first time turned it on, LED lights ok, nothing burnt yet so far, but I'm not able to adjust the current. I have 0.09V reading across R5, and turning the R4 does not make any difference. At this stage, output measurement is 387V, it’s almost exactly the same as raw DC input after CRC. Turning R11 does not make any difference at all.
I also noticed the resistance across Drain and Source on DN2540 is only about 11ohm, ( I have some new DN2540 and they are the same ) . It looks the CCS part is not working. What could be the problem?
Thanks in advance.
The SSHV2 I'm building is for my preamp. The preamp runs 290V@5ma per channel. The DC after rectifier is 400V. I used CRC to drop the voltage to around 350V. I used 80K/5W resistor as dummy load for the reg. I understand I should set the R4 to get 0.25V reading on R5 ( 20ma+5ma ) , then adjust R11 to get 290V for output.
My problem is right now, at first time turned it on, LED lights ok, nothing burnt yet so far, but I'm not able to adjust the current. I have 0.09V reading across R5, and turning the R4 does not make any difference. At this stage, output measurement is 387V, it’s almost exactly the same as raw DC input after CRC. Turning R11 does not make any difference at all.
I also noticed the resistance across Drain and Source on DN2540 is only about 11ohm, ( I have some new DN2540 and they are the same ) . It looks the CCS part is not working. What could be the problem?
Thanks in advance.
Update: I made mistake with both KSA1381 installed back to front.
Now I can set the CCS to 25ma no problem.
However, I can't adjust the output to anywhere higher than 150V.
Now I can set the CCS to 25ma no problem.
However, I can't adjust the output to anywhere higher than 150V.
Update 2 : I think the problem being not able to increase the output is the IDSS. Uisng a 9V battery to test the IDSS, it was 2.4ma. I found one with 3.7ma amoung 50pcs 2SK117 and now I can adjust the output to 280V but no higher than that. I will try some other JFET from different source.