Correction .. The one that creeps upwards does so at .001V increments, not .01. The creep rate seems to have slowed down after 10 or so minutes, though there is still some creep.
Set it after 15 mins warm up. The Jfets and Mosfets have some temperature coefficient, I don't think you got a particular problem in your build unless its set to unusually hot temp regarding its sinking. You could put a minisink on the Vref CCS JFET, or use a Zener reference instead of a resistor/trimmer (Norton Vref). There is provision on the see text PCB area. If get go precision and no warm up slide are paramount in some application. The Zener should match the amount of voltage seen on the resistive part only. For the one going lower short one of its two Vref Leds first and test again.
Hi Salas,
Thanks for the reply. Ok, the 5V BiB supply is off my worry list, the one that creeps up. After 20 mins or so it stabilized to my set voltage, so that's that.
It's the other one that still worries me (the 5.2V). After climbing up to the set voltage it decreases .001V every second (unlike the 5V supply whose change occurred every 5s or so, and slowed down after a while). It's come down to 4.5V from 5.2V, with no signs of stopping before I shut it down.
About the heatsink, the one for Q101 is oversized, while the one for Q106 is slightly undersized (24 deg/W as opposed to the 20.03 deg/W recommended by the calc), but I'm thinking that's still within the safety margin. I do have better sinks on the way.
Any other suggestions? I have access to an o-scope if that can help (though I really haven't had a chance to learn how to use it yet)
Thanks for the reply. Ok, the 5V BiB supply is off my worry list, the one that creeps up. After 20 mins or so it stabilized to my set voltage, so that's that.
It's the other one that still worries me (the 5.2V). After climbing up to the set voltage it decreases .001V every second (unlike the 5V supply whose change occurred every 5s or so, and slowed down after a while). It's come down to 4.5V from 5.2V, with no signs of stopping before I shut it down.
About the heatsink, the one for Q101 is oversized, while the one for Q106 is slightly undersized (24 deg/W as opposed to the 20.03 deg/W recommended by the calc), but I'm thinking that's still within the safety margin. I do have better sinks on the way.
Any other suggestions? I have access to an o-scope if that can help (though I really haven't had a chance to learn how to use it yet)
Short one Led in the declining one as I wrote. From the two in the Vref area. See what happens, let me know. Maybe you chanced on some sample parts tempco combination that slides negative. The other one you made gains a bit with warm up so its a local thing.
Just tried it out ... First time turning it on, there's a measure of ramp up, after which it starts its decline.
Relatively quick turn on / off, no more ramp up, just a decline in voltage.
There was a time or too where the voltage seemed to stabilize after dropping quite a bit, after which I tried using the trimmer to set it as high as I could. Voltage would continue dropping again after that though, dropping past the voltage point which had previously seemed stable.
Relatively quick turn on / off, no more ramp up, just a decline in voltage.
There was a time or too where the voltage seemed to stabilize after dropping quite a bit, after which I tried using the trimmer to set it as high as I could. Voltage would continue dropping again after that though, dropping past the voltage point which had previously seemed stable.
Do you have another trimmer or a normal resistor about the value where the trimmer sets? Just to substitute and make sure it ain't a trimmer thing.
Hmm ... I don't have any 1k trimmers left. I do have some 5ks though. Will that work as well?
Ok, removed the trimmer, soldered a jumper in its place. Removed the jumper from the resistor position and put in a 1k resistor in its place.
Still the same behaviour, after initial ramp up, it shows continuous voltage drop.
Still the same behaviour, after initial ramp up, it shows continuous voltage drop.
Used a red LED, yes. Though that LED was still shorted as per your first suggestion when I tried it out just now.
My previous Vref arrangement was : 1 red led, 1k trimmer, all other items jumpered.
About the CCS extra current, there's 150ma more current than the target load requires.
About the CCS extra current, there's 150ma more current than the target load requires.
Thanks for the suggestion Salas. 🙂 Will try it out in a couple of hours or so after some IRL stuff.
If it will not come to terms, use 2 red leds, 1 diode and one small resistor to construct the 5.2V. Diode in reverse orientation to the printed Zener symbol. There is 0.6V Vbe residual there, so you build from 0.6V up.
P.S. If its not an oscillation making it unstable, bcs the other one build doing different things is weird, the above Vref being mostly active should do it.
P.S. If its not an oscillation making it unstable, bcs the other one build doing different things is weird, the above Vref being mostly active should do it.
Chalk this one down to a noobish mistake.
After switching to an active Vref configuration, voltage was still sliding. It was at this point that I did a visual comparison between the wonky supply and the one that stabilized.
Turns out I had the polarities reversed on both the Vref cap and C107. Unsoldered the offending caps, resoldered them minding the correct polarity, and voila! Stable Vout on load resistor.
Hooking up both supplies, there is still a gradual voltage drop, but it's now in the rate of .001V after every minute or so. Could this just be a case of undersized trafo? Have a single 30VA toroid powering both supplies.
Active vref config : one red LED, one green LED, one 2.7V diode reversed to give maybe 1V? All other items jumpered. Came around to 5.23V initially, currently sitting at 5.18V after more than an hour of playing.
After switching to an active Vref configuration, voltage was still sliding. It was at this point that I did a visual comparison between the wonky supply and the one that stabilized.
Turns out I had the polarities reversed on both the Vref cap and C107. Unsoldered the offending caps, resoldered them minding the correct polarity, and voila! Stable Vout on load resistor.
Hooking up both supplies, there is still a gradual voltage drop, but it's now in the rate of .001V after every minute or so. Could this just be a case of undersized trafo? Have a single 30VA toroid powering both supplies.
Active vref config : one red LED, one green LED, one 2.7V diode reversed to give maybe 1V? All other items jumpered. Came around to 5.23V initially, currently sitting at 5.18V after more than an hour of playing.
Last edited:
Glad you found the cause, it was simply too weird to happen in one build only.
It has to have some mini slide as the LEDs and the BJT come to temperature. Being a negative one is good for not burning something digital if voltage spec is on edge. Just fix it for when hot. Does it fall below the useful low ever?
It has to have some mini slide as the LEDs and the BJT come to temperature. Being a negative one is good for not burning something digital if voltage spec is on edge. Just fix it for when hot. Does it fall below the useful low ever?
- Status
- Not open for further replies.