I will say not only the first 100 but all, at 6k and 3k for the kit they should come to with a bag of gourmet coffee and a 6 pack of Rare Vos Amber Ale. >:)
If he said it then I bet he will do it.
They can use the same set up system they used for the NPC tests. They actually moved them through quite quickly.
With a dedicated set up team they could probably do 15-20 per day.
We are so dialed into the sonic signature that 10-15 minutes each is more than enough to spot any anomaly.
The units need an overnight warm up from a cold boot to be at the optimal operating zone and don’t forget the final SQ signature will occur after a few weeks of burn in at your home.
We should stop speculating: there’s no reason to assume anything has changed since Paul first said they’d listen to the first 100 in the reference system. I did send a message to Paul about this thread, so I expect he’ll answer when he has time.
FWIW It’s very important to listen to a number of the initial units. We need to make sure we are testing all of the important items on each board and each system as they’re manufactured. As time goes on we’ll both refine those tests and gain more confidence in them.
Also it doesn’t take a lot of listening to validate that things sound good. The board tests and system tests have already checked the function of the device (including verifying levels, etc. on a scope and cheapo speakers). The listening tests are just to verify that nothing is slipping thru the cracks: we sure don’t want a black eye on the early systems.
I’m sure that the first few units will take some time, but later there’ll be proportionally more time taken up moving and hooking up the DACs…
- 1 ted
Al
@tedsmith
Thanks for the info and sending a message to Paul for further clarification on the QA process for the DS. Like Oddeophile, this is important to me since I also prepaid for one of the first DS units to be shipped out.
Bob
The listening tests are just to verify that nothing is slipping thru the cracks: we sure don't want a black eye on the early systems.
Ted, how will this be insured for the DIY kits?
Hi guys. Sorry, Terri and I took Saturday to go cross country skiing before all the snow’s gone. It wasn’t a beautiful sunny day but it was good to get out of the house and not stare at a computer screen anymore.
Indeed, I said we’d listen to the first 100 in Music Room One and we will. No, I am not going to sit and play each one for an hour, but to be honest I’ve gotten pretty damned good at picking out DS vs. something less than DS. Yes, its sound is that distinctive. The sales people tried to trick me the other day in the listening room. Using one of the tracks I know pretty well, they pulled me in and asked “PWD or DS?” It didn’t take me but about 10 seconds to tell them it was obviously the PWD. There’s just a flatness to the sound with every other DAC I’ve had in the room. I can spot a DS in about 30 seconds to see if it’s got the magic or not.
So yes, they will wheel in the “patients” one by one and I’ll quickly verify they have the magic before sending them out. What we won’t have time to do is check every input, etc. but those checks are already made in FST and each unit goes through a barrage of tests before leaving or ever getting into the room for auditioning.
As for kits, there’s no way to do that and so they will not be listened tested. However, I suspect they won’t need to as I am guessing there’s really no need to listen to all 100 of the first DS. If my guess is correct, all 100 will fly through the process. I just want to make sure DS magic is there, repeatable, and without much in the way of variations. So far, of the 10 or so we’ve built that’s been the case.
Ted’s done a magnificent job of making sure any of the obvious points of variability potential are extinguished in the design and now it’s simply a matter of proving that. If the 100 fly out the door there’s no reason on earth the kits that follow won’t sound identical.
Thanks Paul. What sort of testing can/will be done on the kits?
I guess the most important result will be if you report back that some units did NOT pass the test…
Thanks Paul. What sort of testing can/will be done on the kits?
The DS board test procedure started from the PWD board test procedure: power supply levels, all of the inputs checked, all of the outputs checked, all of the UI features, etc. Some of this is simpler than the PWD since there's no filter selection, etc.
To test a digital card we pair it with a known good analog card and visa versa.
Some of the new tests include checking bit perfect reception, checking that the VCXO has the expected pull range and dynamics.
Testing the FPGA is in some sense simple, it's real time audio code. Most possible problems will lead to audio artifacts. Just testing the UI while listening will exercise essentially all of the FPGA.
I've taken enough boards in and out of a chassis that I'm comfortable that that is unlikely to cause problems, at least if you wear a static strap, have the power off, double check the cables, etc. :)
We've not yet encountered a board set that passes the above but doesn't sound right. If we ever do we'll modify the board tests.
About all that's not tested is how it sounds in your system.
Come now Ted. That’s not the correlation being suggested. More along the lines of how it will be tested to insure that when we do get it, it will sound the way it’s supposed to once installed. 
Boy I needed to look at the specific smiley face to see if you were serious 
(I’m a literal minded cuss: if you were serious I don’t know what correlation you are talking about.)
Obviously it would not be practical to install the kits, listen, and then uninstall for delivery.
If the 100 fully built and listen tested DSs, with randomly selected boards, all pass the “magic” test, then it is reasonable to expect the rest to so so as well. They are, as mentioned, already bench tested.
Add to that the fact that this is a “full gut” upgrade there are no previous parts to be compatible with except the front panel screen. I think it should be a safe enough process.
Gordon said: Add to that the fact that this is a "full gut" upgrade there are no previous parts to be compatible with except the front panel screen. I think it should be a safe enough process.
Was it (will it be) tested , by the way? And sound-checked afterwards? At least for a couple of kits and on a couple of old PWD from different versions. Not mk1 vs mk2, but different LCD screens.
Just curious, since the firmware was changed to be compatible with the new LCD and that changed the sound.
@alekz
As stated by Paul and Ted, ALL 150 boards and are bench tested including swapping between power and audio boards.
They then build 100 DSs from a random selection of the 150 boards. I think an almost 70% listen for the “magic” in this case should be adequate. At this juncture it is more about anything in the assembly process that could cause a deviance.
Hi guys. Sorry, Terri and I took Saturday to go cross country skiing before all the snow's gone. It wasn't a beautiful sunny day but it was good to get out of the house and not stare at a computer screen anymore.
Indeed, I said we'd listen to the first 100 in Music Room One and we will. No, I am not going to sit and play each one for an hour, but to be honest I've gotten pretty damned good at picking out DS vs. something less than DS. Yes, its sound is that distinctive. The sales people tried to trick me the other day in the listening room. Using one of the tracks I know pretty well, they pulled me in and asked "PWD or DS?" It didn't take me but about 10 seconds to tell them it was obviously the PWD. There's just a flatness to the sound with every other DAC I've had in the room. I can spot a DS in about 30 seconds to see if it's got the magic or not.
So yes, they will wheel in the "patients" one by one and I'll quickly verify they have the magic before sending them out. What we won't have time to do is check every input, etc. but those checks are already made in FST and each unit goes through a barrage of tests before leaving or ever getting into the room for auditioning.
As for kits, there's no way to do that and so they will not be listened tested. However, I suspect they won't need to as I am guessing there's really no need to listen to all 100 of the first DS. If my guess is correct, all 100 will fly through the process. I just want to make sure DS magic is there, repeatable, and without much in the way of variations. So far, of the 10 or so we've built that's been the case.
Ted's done a magnificent job of making sure any of the obvious points of variability potential are extinguished in the design and now it's simply a matter of proving that. If the 100 fly out the door there's no reason on earth the kits that follow won't sound identical.</blockquote
Bingo! Thanks to Paul and Ted confirming. Now those neurotic types like me can sleep well over the next couple of weeks with visions of DS dancing in our heads... Ted also hit the nail on the head with what was floating in my pea brain I was poorly attempting to state.
Well done. Well done, indeed.
@alekz
As stated by Paul and Ted, ALL 150 boards and are bench tested including swapping between power and audio boards.
They then build 100 DSs from a random selection of the 150 boards. I think an almost 70% listen for the "magic" in this case should be adequate. At this juncture it is more about anything in the assembly process that could cause a deviance.
My question was more about the compatibility with different screens.
Was it (will it be) tested , by the way? And sound-checked afterwards? At least for a couple of kits and on a couple of old PWD from different versions. ...
Just curious, since the firmware was changed to be compatible with the new LCD and that changed the sound.
I personally have put many sets of boards in my box (very old PWD chassis) and also multiple PWD mk II chassis and listened. I've seen different reactions on the LCD screens to partial loads, etc. so I know that I've experienced multiple LCD's personally. The DS cards are much less susceptible to changes in the PIC code. (For example I run all control signals as slowly as possible instead of as quickly as possible. I also have 2k resistors in series with all control signals from the PIC. The FPGA doesn't need any baby sitting to do it's job so the PIC just polls status now and then.) To my knowledge there's been no audible difference with PIC code changes or levels of LCDs with DS cards.
Thanks Ted, exactly what I wanted to know
@admin
Thanks for the explaination. It’s most appreciated on your day off
One question though, once the first 100 have been listened to and sent out what is the QA process that PS Audio uses from that point on, not only the DS DACs, but your other products such as the power plants?
Bob