Salas SSLV1.3 UltraBiB shunt regulator

[lasercut]

Q4 vbe 0.59

R1 voltage drop is also accurate (0.6/5.5=0.109)

btw thanks for bearing with me on this issue

 

[Salas]

This constant current setting isn't enough to allow a nice 100mA spare even if the load will be light. Do you test loaded or unloaded?

 

[lasercut]

Unloaded.
Guide suggested 50ma spare is ok, at least for intial unloaded testing?

 

[lasercut]

I am not sure if there is a problem or voltage is just too high?
LED drop + slightly higher Vbe should add to 5.1 but its gives 5.5 so that makes think there is.
5.5 is within limits of soekris so not really an issue i.e no need to change LED if no other problem.

 

[Salas]

R7 should indicate appreciable mA though, around 2. Else its not right.

 

[lasercut]

Is increasing R1 current your next suggestion?

I have another kit for positive reg so I can replace any of the transistors

 

[Salas]

Yes, maybe not enough through M2 thus the lower VGS than 3V

 

[Salas]

Spare a hundred mA when deducting the expected mA load from the CCS limit. Its a safer number. In this case at least. Also compare R7 mV drops across sections. Set the DMM at DC mV.

 

[lasercut]

Def something wrong, increasing current to 175mA caused minimum output voltage to increase to 6V...

 

[Salas]

Weird things first time I see in this thread. Does J2 do anything? Pulls? R7 should show about that.

 

[lasercut]

I need to leave it for tonight , thanks for help so far

 

[Salas]

Me too, see ya

 

[lasercut]

on R7 there is 70mV now.

This is with:
15V input
175mA R1 current
min 6.6V output

 

[Salas]

70mV is too little. We want 300mV or more. Is it the same on the negative section?

P.S. Parallel a small value resistor to R7 like 47 Ohm. That will make R7 40 Ohm. How much current now? (R7Vdrop/40)

 

[lasercut]

neg is around 350mv, thats still using 9v input and previous lower R1 current aswell.

R7 voltage drop was same with lower resistor (.073/37 = 0.0009A

 

[Salas]

Replace J2 if you have a spare. Then Q2 Q3 if nothing changes. Ultimately replace led1 with a normal silicon diode. Cathode at same direction as the led's.

P.S. Note M2's Vgs with each experiment.

 

[lasercut]

Replaced J2, have 400mV on R7 and output it down to 4.8V

I have nice 6VAC toroid that I used for TPS7A regs which I would like to use for closest comparison between the ultraBib
Guides mention 1.7 Vin-Vout is min for <200mA, it would be 4.5 Vin-Vout with this trafo.
Do you think it would be unfair to compare UltraBib to LDOs at 4.5V voltage difference?

 

[Salas]

Good news. J2 got hit by soldering or static? Was it just a random dud? Who knows.

Ultrabib wants >5V in-out for faster CCS PSRR response because M1 becomes less capacitive with more voltage across it. Not that it doesn't work with less. The guide has some details on that explaining the absolute necessary voltages for some current thresholds if compromising quickness. Fairest comparison would be with ample voltage across M1. Benefit starts at 5V and stops at 20V in-out where most Mosfets capacitance curves stop losing pF and flatten.

Anyways, no matter the configuration details, Ubib will always come across different to LDOs subjectively. Its a discrete shunt and that is its nature. Different. You will decide what synergy you will prefer. But give its electrolytics a chance to come up with some use first. They need little conditioning after storage time in warehouses.

 

[lasercut]

Up to 20V? that's good because I only have 15VAC trafo on hand to replace it with so at least the extra dissipation will not go to waste.

Will let the regs run for a while, soekris could probably benefit from some warm up time too.

 

[Salas]

Alright, enjoy listening, and you let us know what you will ultimately prefer in your system for your taste. Ubib or LDO. Don't use too thin output cables by the way.