SSR03 - TO92 transistors, hFE requirements

[tk-don]

Hi P-A.

I was about to mount all 6 TO92 type transistors in my SSR03, but I just noticed that in your BOM spreadsheet you write hFE>420, or more precisely for the PNP: "Transistor, PNP, 65V, 100mA, 625mW, 300MHz, Hfe > 420, TO92".
The voltage rating, collector current and power dissipation and GBW product I guess are just the usual specs of the BC556B, but where does the DC current gain requirement/performance come from? I don't see any manufacturer claiming it for the BC556B regardless of collector current.
The 556B I got measures 370 at Ic = 2mA and 5V. But since it is below the 420 I wanted to ask: Is it OK to use the BC556B/BC546B and is the hFE of >420 of irrelevance here ?

Thanks!

Regards,
Troels

[peranders]

The "B" selection will also work it only limits the max current a bit.

[tk-don]

Fair enough. Since it turns out the 546B has an even lower gain (317) and that I've got a ton of reasonably matched BC550C/BC560C: Would these be possible to use instead without any modification to other components?

[peranders]

Yes they will also work.

[tk-don]

I thought I'd just re'use this thread.
Got everything soldered.
The configuration is +/-15V, LM317/337 pre-regulators, AD825 opamps, TL431 (0.4%) voltage references soldered into the positions with the bold outline. I'm using a 20k multiturn pots in P3 & P4 to fine tune the reference - R39 & R40 not installed.

Testing with a bench power supply (with current limit of 100 mA) at 22V at the connector at the X1 conector shows that the output voltage is close to +/-15V, the LEDs light up and becomes dimmer when the input voltage decreases to a little below 19V. So far so good!

The problem occurs with loading. R49 & R50 are 1 ohm (3W resistors), so the current limitation should be 0.65A per voltage rail according to your notes, right? I have tested it with an electronic adjustable load. Both rails behave the same:

Output voltage, load current:
14.93 V, 0.28A
14.75 V, 0.35A
13.92 V, 0,43A (LED is very dim)

What also happens while increasing the load is that R11 & R12 - which are 5.6 ohm - heats up, and with time becomes a little too hot to touch at 0.43A loading, not something that would damage the component, but I suppose I should have gone with a lower resistance.

I have also measured the output voltage of the AD825 at pin 6 which is 8.1V at no load, then rises at remains steady at 13.6V with 0.35A load, and 12.6V at 0.43A.
All BC550C and BC560C I use are matched to a hfe of 598+/-3 at 25 degrees.

I can attach an image of the component side of the PCB, but I think it's best that I measure the voltage at all positions to find the problem. Do you have a list/hint as to what the voltages should approximately be at various points of the circuit, like for the QSXM2?  :-)

Thanks for any kind of hints in advance.

[peranders]

You must have 4.5 volts more at least at the LM317 input pin, do you have that? = 19-20 volts at least.

[tk-don]

I will measure this for sure again tomorrow, but I was pretty sure it was measured to 20.3V or something (without load). Anyways, I'll try with a high supply voltage (25V).
Does the no-load output from the AD825 of 8.1V seem reasonable? What is worrying a bit is that it increases in order to compensate the dropping output voltage, but then decreases to 12.6V.
Anyways, I'll write back later  :-)

[tk-don]

Success!  :-)
Indeed the key was that the voltage at the output of the LM317/337 depends on the load:

24V supply:
LM317 input, load current:
21.09V, 0.28A
20.46V, 0.37A
20.09V, 0.43A

Very similar for LM337
All resulted in an output of 15V, within a few mV deviation - this may also have been affected by the wires going to the adjustable load. I didn't bother to measure any other locations of the spower supply. But the efficiency is about 55% (11.6W drawn -> 6.4W output) and seem free of noise with "square wave" load transients undetectable in the voltage waveform.

The 5.6 Ohms in position R11 and R12 get quite uncomfortably hot to touch at 0.5A load. The voltage drop across it is about 3V, so  0.54A, or approx 1.6W. Providing the calculation is correct, there's no problem, but would there be any benefits of increasing the value - you generally advice that this resistance should be as low as possible?

Thanks again!

[peranders]

The series resistors should be chosen after application. Big current => lower the values.