sonic difference between linear and switch mode PSU?

[Rochey]

PS: As i recall we had hired one full time switching supply expert and kept him busy on multiple projects. between mixers and power amp groups.

SMPS is quite a black art... experts should be kept in dark rooms, and fed through letterboxes... (kind of like mushrooms)

 

[JohnRoberts]

PS: As i recall we had hired one full time switching supply expert and kept him busy on multiple projects. between mixers and power amp groups.

SMPS is quite a black art... experts should be kept in dark rooms, and fed through letterboxes... (kind of like mushrooms)

The laws of physics involved are finite, like hair care.... ;D

However there are a number of different considerations that take time to get up to speed about. Expertise comes from already making all the common beginner mistakes before.

I was apprehensive about switchers years ago, and not an expert now, but I have used them sparingly and had to redesign one troubled complex design I inherited (mostly by falling back to a simpler approach).

FWIW my very first technician gig back in the late '60s was working on a DC to DC switching PS for a navy rescue vessel. That PS was designed by Raytheon and being debugged by engineers at MIT Instrumentation Lab. I was a junior tech at the beginning of the project tasked with replacing parts the junior engineer blew up. By the end of the project I was blowing up parts all by myself.  8) Back in the 60s it was a relatively new technology. None of the canned IC solutions like today. As I recall the Raytheon engineer tricked a then brand new analog voltage regulator IC (LM100 ?)  into switching. 


Regarding switch mode engineers, reflecting back on our hired gun at peavey, he was a little lacking in social skills. I literally had him removed from one product meeting when he wouldn't stop arguing about something that bothered him. I was actually doing him a favor, while he didn't appreciate it at the time (you don't pick a fight with the guy running the meeting.. I had my own Roberts rules of order.  ).

JR

 

[Andy Peters]

I know that SSL had a huge filtering box that was part of the SMPSU on their 9000 series consoles.  I found one mention of that unit, apparently filled with a bunch of L and C components, and that filter box is 9 RU in size and weighs 200 KG.

Interesting-- particularly since one of the advantages of SMPS is usually said to be size and cost.

Sounds like the SSL supply used a low-ish switching frequency, hence larger components. And it is powering a huge console -- how many tens of amps? -- so that too requires larger components (to avoid saturating magnetics, etc).

How big was the linear supply that the switcher replaced?

-a

 

[Andy Peters]

I have used 48V switchers for phantom power near transformer coupled preamps, with no discernible noise contribution.  In practice many have lower output noise than a comparable linear, with the noise performance best at lower loads, opposite of many linears.

I used a TI switcher to give me 48 V phantom from 5 V USB and you can't tell that there's a switcher in the circuit.

You just have to follow the rules -- mind your loop currents and your switcher return paths, keep that all away from the preamp and other analog, etc etc.

-a

 

[Rochey]

Andy, which one?

 

[JohnRoberts]

Back in the early '90s when Peavey developed twin consoles separated at birth to serve Live and Recording markets with different versions using common components and tooling. The lead engineer for the live console used a switching supply for his 48V to help keep his PS smaller lighter. He used conventional supplies for the other higher current rails so I don't know how much size/weight he actually saved. IIRC his PS was 2U while mine was 3U. i think he also used a trick transformer, while mine was old school big-heavy EI core.

Since I was lead engineer for the recording console my 48V rail was very conventional, extra winding on my already custom transformer. I had to support 36 inputs so a few amps. I saw no reason to mess with switching technology for a power supply that should get lifted very infrequently. For the live console making the whole power supply a switcher would make some sense but I suspect he chickened out, or had an early set back (I don't know and never asked).

As chance would have it I still ended up with a small digital midi controller embedded in my master section so i wasn't completely free of digital/switching poop, but it's easily managed if you know it's there (put it inside a metal box and filter any lines coming or going).

JR

 

[soapfoot]

I know that SSL had a huge filtering box that was part of the SMPSU on their 9000 series consoles.  I found one mention of that unit, apparently filled with a bunch of L and C components, and that filter box is 9 RU in size and weighs 200 KG.

Interesting-- particularly since one of the advantages of SMPS is usually said to be size and cost.

Sounds like the SSL supply used a low-ish switching frequency, hence larger components. And it is powering a huge console -- how many tens of amps? -- so that too requires larger components (to avoid saturating magnetics, etc).

How big was the linear supply that the switcher replaced?

-a

The supply for our 4040E is 3 or 4 RU (can't remember offhand).  I think the 4056 and larger 4k series had two of those supplies.

 

[Andy Peters]

Andy, which one?

TPS61081. There's an app note for the part which shows a 5V-to-48V converter. Put a doubler on the 24V output and off you go.

-a

PS: I'd provide a link to the app note but the TI web site seems dead for some reason.

 

[PRR]

> a link to the app note

http://www.ti.com/lit/ds/symlink/tps61081.pdf

Fig 26 pg 18 shows the 5V-50V scheme with values. Shave a few dozen Kohms off R1 if you like 48V.

 

[Brian Roth]

I know that SSL had a huge filtering box that was part of the SMPSU on their 9000 series consoles.  I found one mention of that unit, apparently filled with a bunch of L and C components, and that filter box is 9 RU in size and weighs 200 KG.

Interesting-- particularly since one of the advantages of SMPS is usually said to be size and cost.

Sounds like the SSL supply used a low-ish switching frequency, hence larger components. And it is powering a huge console -- how many tens of amps? -- so that too requires larger components (to avoid saturating magnetics, etc).

How big was the linear supply that the switcher replaced?

-a

As I recall, SSL shipped that switcher and filter bank as stock for the 9000 desks.

Bri

 

[Andy Peters]

> a link to the app note

http://www.ti.com/lit/ds/symlink/tps61081.pdf

Fig 26 pg 18 shows the 5V-50V scheme with values. Shave a few dozen Kohms off R1 if you like 48V.

that's what I did

After building it, I later found this app note, which shows all work: http://www.ti.com/lit/an/slva279/slva279.pdf

-a