English Amiga Board

English Amiga Board (http://eab.abime.net/index.php)
-   support.Hardware (http://eab.abime.net/forumdisplay.php?f=20)
-   -   Jens Schönfeld - Amiga PSU Seminar (Revision Party 2020) (http://eab.abime.net/showthread.php?t=101682)

kriz 11 April 2020 19:24

Jens Schönfeld - Amiga PSU Seminar (Revision Party 2020)
 
Earlier today Jens had a very nice seminar of Amiga and its PSU.. Comparing the old and new alternatives and how good they really are for updated Amigas. Topics:
  • Existing solutions
  • Original Commodore spec
  • Load Performance, observations
  • Strategies against voltage drop
  • Milestone: New specification
  • Basic principle of DC-DC conversion
  • Compensation circuit & performance

Check the recording here: https://www.youtube.com/watch?v=nsKnbcBnay8

vroom6sri 11 April 2020 21:16

Methinks I might be ordering one of their 060 accelerators and one of these new power supplies when they have both ready. Thanks, Jens!

Pyromania 11 April 2020 21:33

Cool!

Pixelfill 11 April 2020 21:48

Sounds good from a number of perspectives.
1st and foremost Jens has found his Mojo, and motivation again.

kriz 12 April 2020 01:42

Yeah this is very nice, also interesting to see those Meanwell and Cisco PSU`s that a lot are using now (me included) are not that well suited for a upgraded system.. I will be looking out for his new PSU also ..

Mathesar 12 April 2020 09:50

The voltage drop compensation circuit is basically a negative resistance. I have built something similar in my previous job to compensate the changing resistance over time for a "liquid" accelerometer.
I wonder how that circuit handles dynamic loads; things like current steps, startup behaviour, etc. Messing with feedback in switched dc/dc converters is always scary, especially with another switcher feeding it!

grond 12 April 2020 10:23

Quote:

Originally Posted by kriz (Post 1391242)
interesting to see those Meanwell and Cisco PSU`s that a lot are using now (me included) are not that well suited for a upgraded system..

He's been telling this BS for a long time. I think he invented the word "current ripple" specifically for it.

Jope 12 April 2020 10:35

Quote:

Originally Posted by grond (Post 1391293)
He's been telling this BS for a long time. I think he invented the word "current ripple" specifically for it.

You don't have to like Jens, but please don't attempt to discredit him by alleging that he invented himself the fact that ripple in power supplies is a bad thing.

https://en.wikipedia.org/wiki/Ripple_(electrical)

Mathesar 12 April 2020 11:36

Quote:

Originally Posted by grond (Post 1391293)
He's been telling this BS for a long time. I think he invented the word "current ripple" specifically for it.

Ripple can be bad, but it depends on the application. For the Amiga, I don't think it is really critical. My own experience with 060's shows that they can handle a lot of ripple before things get really bad.
That being said, a good stable power supply is important and a bad power supply can cause a lot of trouble! Kudos for Jens trying to address this.

grond 12 April 2020 16:34

Quote:

Originally Posted by Jope (Post 1391296)
please don't attempt to discredit him by alleging that he invented himself the fact that ripple in power supplies is a bad thing.

https://en.wikipedia.org/wiki/Ripple_(electrical)

Your link talks about a voltage ripple, not a current. In any case he presents it as some magical intrinsic quality that he discovered when it is just power supply noise.

Daedalus 12 April 2020 17:10

Current ripple will always be present with any voltage ripple. You can't increase or decrease the voltage applied to a circuit and expect the current to remain static, so I don't really get why such a big deal is being made of it. It's just another facet of the same interaction. It's also been a specification important for capacitors used for smoothing ripple, as applying ripple current beyond a capacitor's rating will cause it to heat up and degrade over time.

As for the modular PSUs, some have higher ripple specifications than others, even from the same line of products. The Mean Well RPT-60B, which is commonly used in all those small replacement PSUs currently available, has a higher ripple rating than a PT-65B for example, so it's really a case of picking an appropriate model for the application. That said, ripple ratings are based on a PSU under maximum load, and won't be the ripple seen under typical use since you would never want to use a PSU at the limit of its capacity.

grond 12 April 2020 17:48

Quote:

Originally Posted by Daedalus (Post 1391395)
You can't increase or decrease the voltage applied to a circuit and expect the current to remain static


Yes, you can. This magic property is called "power supply rejection ratio" and describes how immune a circuit is to noise present on the supply voltage. Also, there are bandgap reference circuits commonly used on semiconductor chips which produce a reference voltage totally independent of the actual supply voltage.



Talking about the dangers of "current ripple" is pretty much like the dangers of dihydrigenoxide. It may be dangerous but usually it is not. It's not like Meanwell don't know how to build power supplies...

Daedalus 12 April 2020 18:28

Quote:

Originally Posted by grond (Post 1391408)
Yes, you can. This magic property is called "power supply rejection ratio" and describes how immune a circuit is to noise present on the supply voltage. Also, there are bandgap reference circuits commonly used on semiconductor chips which produce a reference voltage totally independent of the actual supply voltage.

That energy goes somewhere though, it's not magic. We're not talking about theoretical, ideal circuits here, but the real, imperfect, complex loads without regulation that are Amiga motherboards.

Quote:

Talking about the dangers of "current ripple" is pretty much like the dangers of dihydrigenoxide. It may be dangerous but usually it is not. It's not like Meanwell don't know how to build power supplies...
Mentioning it was just to illustrate that it's a well know and documented phenomenon taken into account in designs, and not something that should be new to any engineer. I don't know why you seem to think I was scaremongering or anything, I know and understand it well. It was also nothing to do with Mean Well supplies, or any other PSU. Of course they know how to make supplies, and they're rather good at it too. But their products cover a wide range of applications, and the specifications (including ripple) differ accordingly.

grond 12 April 2020 19:57

Quote:

Originally Posted by Daedalus (Post 1391418)
That energy goes somewhere though, it's not magic.

It's not energy, it's a potential difference. One that may produce a proportional current flow but usually doesn't because of the non-linearity of circuits such as semiconductor circuits. That's precisely the reason why digital circuits are essentially overdriven analogue amplifiers: they operate in non-linear mode and thus suffer much less from power supply noise.

Mathesar 12 April 2020 22:05

I don't hear Jens using the words "current ripple" in this video.
What I did find interesting in his video is how well the old PSU's held up in the ripple test. Old "dried up" electrolytic capacitors have increased internal resistance called ESR. Increased ESR causes increased ripple. I have actually tested some 30+ years capacitors in my A500 and the ESR was fine.
I think that the stories about bad electrolytics are often overrated. I still use them quite a lot in the circuits I design.
A lot depends on the manufacturer and how the caps are used in the circuit.
Stick with the likes of Rubycon, Panasonic, etc. Keep the ripple current low and keep them cool and they will work forever.

Mathesar 12 April 2020 22:11

Quote:

Originally Posted by grond (Post 1391430)
It's not energy, it's a potential difference. One that may produce a proportional current flow but usually doesn't because of the non-linearity of circuits such as semiconductor circuits. That's precisely the reason why digital circuits are essentially overdriven analogue amplifiers: they operate in non-linear mode and thus suffer much less from power supply noise.

Ripple current is more of an issue on the input side of a switched dc/dc converter. A buck converter with a small input cap and a low impedance path to tbe power source will draw a lot of ripple current and emit a ton of interference if not dealt with properly.

Pyromania 13 April 2020 05:15

New kick ass power supplies from Jens soonish?

Hewitson 13 April 2020 07:54

How much will the iComp PSU's be? $3000?

Lord Aga 13 April 2020 09:12

Quote:

Originally Posted by Pyromania (Post 1391529)
New kick ass power supplies from Jens soonish?

Either that or he's just bragging about making the best Amiga PSU in the world for himself only :)

grond 13 April 2020 11:53

Quote:

Originally Posted by Mathesar (Post 1391458)
Ripple current is more of an issue on the input side of a switched dc/dc converter. A buck converter with a small input cap and a low impedance path to tbe power source will draw a lot of ripple current and emit a ton of interference if not dealt with properly.


That's probably why according to the discussion on a1k.org Jens will use an ordinary 3rd party converter for this critical input stage to generate an intermediate DC voltage and only generate the 5V/12V/-12V from that intermediate DC voltage using his own circuitry.


All times are GMT +2. The time now is 17:08.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, vBulletin Solutions Inc.

Page generated in 0.05242 seconds with 11 queries