UDP3: Umbilical Cable Driver for Pearl 3 phonostage -- move Pearl 3 PSU into its own, distant, chassis

Joined 2020
Paid Member
Well, almost.

After long delay, here's my UDP3. I'd been waiting for Gianluca's nice back panel, a cable gland, and a free weekend afternoon. Also I prefer powering up stuff off of the mains only when all the fuses & shrink wrap are in place, so.

Unloaded, the rail LEDs lit & there was 19.23V between positive rail and ground, and -19.27V between negative rail and ground. With 98R7 of power resistor between each rail and ground, there was now 18.86V positive and -18.90V negative, which is just about 190mA of draw per rail. Good. I ran it loaded like that for a few minutes, the power resistors got warm and the heatsinks on Q3 and Q4 got a tiny bit warm.

So then I hooked up the umbilical & completed the box, & for sanity's sake I thought to check the voltage at the other end of the XLR I'm using as an umbilical. But I got clever, and decided to check the potential rail to rail instead of rail to ground, and may or may not have shorted the rails while doing so. Anyway that popped the rail fuses. I don't have spares (I gave my two spares away), so now I get to order more & wait. I'm pretty confident it's the rail fuses that went open because I can still measure 19+V from pin 2 of Q3 and Q4 to ground, and nothing from the ferrite beads after the fuses to ground.

I've done this before, measured from rail to rail in other power supplies and it hasn't been an issue, but then again I wasn't necessarily paying attention to rail currents while doing that, either. I'm surprised this popped both rail fuses, and again maybe I shorted the rails, but, there ya go. Everything fits & looks nice, at least.

BTW if anyone wants a cable gland, I bought a bag of 10 from Amazon & have lots of spares. Will trade for fuses. :geek:


  • IMG_9516.jpg
    752 KB · Views: 143
  • IMG_9517.jpg
    705.9 KB · Views: 143
  • IMG_9518.jpg
    707 KB · Views: 144
Joined 2011
Paid Member
post#1 of this thread said:
... The fuse, a small thru-hole packaged device, blows if/when you mistakenly short UDP3's output(s). It is available on DigiKey and Mouser for about 40 cents per fuse. Because they cost so little, I recommend you buy at least six fuses of each of three different fusing currents. Or more. Now when you get a surprise fuse-blow event during testing, you will have plenty of spares, and they don't cost an arm and a leg.



  • parts_list.png
    20 KB · Views: 66


Joined 2010
Paid Member
Before answering that question, it's extremely important to remember that this project, is made to mate with a very specific project... Pearl 3.

Which has just 10uF in front of the regulators.

Also, do you think that more capacitance will improve it? Or are you thinking of using it to something that already has a bunch of installed capacitance?

Anyway, not trying to talk you out of anything at all, merely bringing some questions up.
Joined 2011
Paid Member
A fuse will blow when someone pushes more current through it, than its "rated current". At the moment my UDP3 board has 250 mA rated current fuses soldered in, and they would blow if someone pushed 300 mA (or more!!) through them.

If I suspected that one or both fuses might be blown, I'd measure the voltage-to-ground (A) upstream of the fuse; and also (B) downstream of the fuse.

I'd also check to see whether the OK-POS and OK-NEG LED's are lit. If a fuse is blown, its LED won't be lit.
Joined 2011
Paid Member
In the UDP3 circuit design, snubbing resistors R1 and R2 are selected to mate with your personal choice of power transformer. The Detailed Parts List does say, in the "Mouser Part Number" column, "QUASIMODO" { spreadsheet cell K4 } . Quasimodo is the diyAudio test jig which helps you find the optimum snubbing resistance. I guess I should have listed "TBD" (to be determined) for the resistor values, on the schematic and the parts list; that might have avoided some confusion. Sorry for that omission! You determine those resistor values using Quasimodo. Or you get lucky and joyfully discover that somebody has already used Quasimodo, on your exact transformer (!!), and has posted their results.

Many diyAudio members have tested many different transformers on their Quasimodo test jigs, and reported the optimized snubber component values to a discussion thread here on the forums: Quasimodo results (ONLY). Have a look at posts # 150 and 151 in that thread. Those are the two transformers which I suspect will be most frequently chosen by UDP3 builders. Don't sweat the precise resistance value too much; anything within plus-or-minus twenty percent will be just fine. Buy a pair of nearby, standard value resistors, at a low price, and sleep soundly. 100 ohms is plenty good enough instead of 98. 56 ohms is plenty good enough instead of 60.

Also please honor member @JeffYoung 's command to ONLY post Quasimodo results in that thread. Got a question? Want to beg somebody for a huge favor? Need an explanation? Keep it out of the results (ONLY) thread. Put it in the Quasimodo big daddy thread instead.
  • Like
Reactions: 1 users
I guess the worst thing that can happen if Quasimodo circuit is a bit off is that it will have little or no effect?
I can never be worse compared to not having it installed at all?
I installed 33R knowing that it will not be perfect for the Toroidy 20VA I will use.
Joined 2003
Paid Member
How do you know that the matching LPs sounded the same when both were new?
It's unlikely, especially if they were made in different production runs, or even in different years.
Many LPs had more than one location where they were recorded. Back in the day, I remember Led Zep albums quality differed from which studio they came from. I have several copies of each as a result. Not that Led Zep is a house favorite, but still a good example of what I am referring to.

There were also a lot of pressings plants around. A lot of pressing matrices and you don't know the quality of the matrice or if the pressing you get it one of the first or last or how many copies the pressing plant makes from a matrice. Then there is also the quality of the vinyl used. Many variables.