It looks like if you need really low frequency rejection and can live with slightly higher noise, then the 400 Zener version should be superior to the standard 002 resistor and cap solution.
The low frequency rejection of the 400 Zener version is absolute terrific.
Thanks for the measurements, great work. 😎
EDIT : But I still think you should try the Exar SPX2431AM, TL431 equivalent. It is so far the lowest noise TL431 or equivalent that I have been able to find(going by datasheet values).
The low frequency rejection of the 400 Zener version is absolute terrific.
Thanks for the measurements, great work. 😎
EDIT : But I still think you should try the Exar SPX2431AM, TL431 equivalent. It is so far the lowest noise TL431 or equivalent that I have been able to find(going by datasheet values).
Last edited:
C3 in #302? I could offer you 1nF... 😉
Not for all the circuits and not this weekend, but I'll see what I can do.
Pass me one and I'll try it. You'll get it back afterwards 😎.
EUR 0,55 is not that expensive, but EUR 20,- for shipping is 🙁. Do you know any source other than mouser?
Maybe their confirmation email sender is not at work anymore, I can't get the registration done 🙁. No samples today.
If this is what you meant, pergo: It's performing almost exactly like #100, except for the noise going down from a couple kHz onwards, being about 5dB lower than #100 at 20kHz. Nothing to rave about 😱. Now I totally want to try that SPX...
If this is what you meant, pergo: It's performing almost exactly like #100, except for the noise going down from a couple kHz onwards, being about 5dB lower than #100 at 20kHz. Nothing to rave about 😱. Now I totally want to try that SPX...
Lasse
I can send you one. I am based in London.
Some dumb questions: DUT = device under test? I'm, trying ti understand the rest of your test rig. The switch is used to step the load? What's the purpose of the series capacitor and diodes?
Regards
Steve
BTW I ordered some PCBs based on the circuit I posted earlier. Awaiting delivery. There will be some spares if anyone is interested.
I can send you one. I am based in London.
Some dumb questions: DUT = device under test? I'm, trying ti understand the rest of your test rig. The switch is used to step the load? What's the purpose of the series capacitor and diodes?
Regards
Steve
BTW I ordered some PCBs based on the circuit I posted earlier. Awaiting delivery. There will be some spares if anyone is interested.
Feeling good about the circuit I posted in post #89 with the TL431 replaced by the SPX431 (and with the simulation posted in #95 suggesting, unless I am missing something, the extra cap, C3 in 302 isn't a good idea)!
Last edited:
Hmmm….
I stuffed my little regulator board today and I have a problem. Instead of reading 5V I am getting 2.45V. I am puzzled as to what I have done wrong.
I'm pretty sure I have the pin assignments for the SPX431 correct. The data sheet has a bottom view of the TO-92 package. I read this to place the cathode at the top when oriented as per the above (with the ref pin at the bottom). If I have this orientation wrong then I would expect the output voltage to be about 10V and change.
Any ideas?
PS: I have 2V at (what I believe to be) the ref of the SPX431 and only 1.2V at the adj pin of the LM317.
I stuffed my little regulator board today and I have a problem. Instead of reading 5V I am getting 2.45V. I am puzzled as to what I have done wrong.
An externally hosted image should be here but it was not working when we last tested it.
I'm pretty sure I have the pin assignments for the SPX431 correct. The data sheet has a bottom view of the TO-92 package. I read this to place the cathode at the top when oriented as per the above (with the ref pin at the bottom). If I have this orientation wrong then I would expect the output voltage to be about 10V and change.
Any ideas?
PS: I have 2V at (what I believe to be) the ref of the SPX431 and only 1.2V at the adj pin of the LM317.
Last edited:
That's it, the graphics are online. For your convenience again:
click here
Yeah, the SPX is actually a lot better than the ET, but not with Mark's circuit (#320). There's some strange low-frequency rumble there which doesn't look nice in the time domain; step response is also impaired.
Now that was quite a bit of work, more than I had imagined. Gotta find a bulky 10 to 15 ohm resistor for pergo's load test now... 😉
I'm planning to test a 7815 for comparison, too, and if I'm really bored over the next weeks I might even examine the benefits of a capacitance multiplier downstream. Maybe even a noise shunt. Getting complicated now, circuit-wise 😉
click here
Yeah, the SPX is actually a lot better than the ET, but not with Mark's circuit (#320). There's some strange low-frequency rumble there which doesn't look nice in the time domain; step response is also impaired.
Now that was quite a bit of work, more than I had imagined. Gotta find a bulky 10 to 15 ohm resistor for pergo's load test now... 😉
I'm planning to test a 7815 for comparison, too, and if I'm really bored over the next weeks I might even examine the benefits of a capacitance multiplier downstream. Maybe even a noise shunt. Getting complicated now, circuit-wise 😉
Did you try to short out R1? Or replace the CCS with a simple resistor (use 243 in this case)?
Everything seems to be correct, be sure to check for soldering errors or even faulty tracks (shorts to the ground layer [don't know your clearance] or discontinuations).
EDIT: NOW! Gate and R1 are supposed to be connected to the Cathode of the 431, not Ref like you did!
Everything seems to be correct, be sure to check for soldering errors or even faulty tracks (shorts to the ground layer [don't know your clearance] or discontinuations).
EDIT: NOW! Gate and R1 are supposed to be connected to the Cathode of the 431, not Ref like you did!
Last edited:
Ugh… My LTspice model was right but the Eagle modelling is wrong. Oh so stupid. Bin it and start again. 🤐 🤐
To go single layer I would have to route the C1 to the input of the LM317 all the way around the board plus rearrange things to deal with R1 to source. Is it such an issue? (I thought the via would be covered by solder mask but it isn't.) The tracks are 1.6764mm. Not enough?
Hopefully this is better:
I haven't done the ground planes or silk screen yet.
(this is becoming the most expensive little DC regulator ever built…I'm certainly learning to pay more attention to detail)
An externally hosted image should be here but it was not working when we last tested it.
I haven't done the ground planes or silk screen yet.
(this is becoming the most expensive little DC regulator ever built…I'm certainly learning to pay more attention to detail)