dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
831
share rss forum feed


shdesigns
Powered By Infinite Improbabilty Drive
Premium
join:2000-12-01
Stone Mountain, GA
Reviews:
·EarthLink
·Comcast
·Atlantic Nexus

Time for an old power supply to retire.

Click for full size
Click for full size
I designed and built this in high school, probably 1977 or so. Made mostly from salvaged parts.

0-25V, 3 amps.

Left knob was a dual pot. Pull and turn to set voltage, push/turn to set current limit.
Right knob had 4 positions for meter, 0-10V, 0-20V, 0-.3A and 0-3A.
Top right switch selects 0-300mA or 0-3A current limit.
Had a low current tracking negative supply and a +5V out.
4 holes around meter were for an Intersil 7107 3.5 digit LED meter that eventually died (eval board.). Replaced with analog.
I think the LED indicated when it had switched to current limit.

I remember when I built it, it was neatly done. Changes, addition, parts burnt out over 30 years left it a mess inside.

Inside 2, 24V transformers in parallel. They are burnt out, probably one developed a short and took out the other. I remember the - supply transformer and board were stolen from something else.

Some things are funny to see. A perf board with 3/16 spaced holes. Control board has two LM741 opamps and and unmarked transistors (RS surplus.) 1/2 watt 10% resistors. Terminal strips (long one in center had rectifiers and main caps but has been removed.)

There is some kind of thermistor on the heatsink. I have no memory of what it did.

Oh, no fuse anywhere. I could have sworn I had one inside. Must have needed it for something....

I worked it hard for years until I built a better one using modern switching supplies (0-30V, 20A.)

WIll trash it. Probably the only things I'll keep is the meter and the current sense resistors.

Funny, I just bought a LM2596 module, 1 3/4" by 3/4" and 1/2" tall that matches its output. Paid About $1.10 each for them.
--
Scott Henion

Embedded Systems Consultant,
SHDesigns home - DIY Welder


aurgathor

join:2002-12-01
Lynnwood, WA
kudos:1
Reviews:
·Frontier Communi..
said by shdesigns:

I remember when I built it, it was neatly done. Changes, addition, parts burnt out over 30 years left it a mess inside.

Hehehehe Been there, done that.

WIll trash it. Probably the only things I'll keep is the meter and the current sense resistors.

Good old analog meters, especially those with mirrors can be quite expensive nowadays. Of course there is always fine old "junk", one just need to find it.

Funny, I just bought a LM2596 module, 1 3/4" by 3/4" and 1/2" tall that matches its output. Paid About $1.10 each for them.

Yeah, many things are just so much better today.
--
.sig


Gabrio

join:2014-08-15
reply to shdesigns
How long it is!


tschmidt
Premium,MVM
join:2000-11-12
Milford, NH
kudos:9
Reviews:
·Hollis Hosting
·G4 Communications
reply to shdesigns
Cool, I don't think I have any electronics left from stuff I built in grade school and high school. My first kit was a Knight VTVM. Saved up all summer in 8th grade to buy it. Built lots of stuff with salvaged TV components. Looks like I'm about 10 years older then you so transistors were still a novelty back then.

Your power supply looks to be wired with salvaged 25-pair key-system telephone cable. I was fascinated by telephones back then. Making it legal to have your own extension took all the fun out of it.

The custom meter calibration is a nice touch.

Funny you mentioned the buck converter. My latest home project is installing a bunch of 12V LED floodlights and PIR sensors around the house and outbuilding. One of them is about 200' feet from the house so voltage drop is a problem. Threw a $1 boost converter to increase 12V to 15.

/tom


shdesigns
Powered By Infinite Improbabilty Drive
Premium
join:2000-12-01
Stone Mountain, GA
Reviews:
·EarthLink
·Comcast
·Atlantic Nexus
Did a bit of tube work. Mostly fixing guitar amps. Had an old Halicrafters SW radio.

Yes, picked up a 1.5' piece of 25-pair somewhere. Was great for breadboard jumpers.

These buck converts work but are noisy. No HF bypass caps. I checked Mouser for the LM2596 and you have to buy 50k before the price gets below $1/ea. Boards were $11 for 10 with free shipping.
--
Scott Henion

Embedded Systems Consultant,
SHDesigns home - DIY Welder


mackey
Premium
join:2007-08-20
kudos:13
said by shdesigns:

I checked Mouser for the LM2596 and you have to buy 50k before the price gets below $1/ea. Boards were $11 for 10 with free shipping.

You don't think that's a legit name brand LM2596 on those boards do you? Hop on over to AliExpress and you can pick up a 20-pack of LM2596 chips for $5.80 shipped.

/M


shdesigns
Powered By Infinite Improbabilty Drive
Premium
join:2000-12-01
Stone Mountain, GA
Yes, assumed counterfeit, stolen or made by slave labor.


mackey
Premium
join:2007-08-20
kudos:13
And perhaps all of the above

public

join:2002-01-19
Santa Clara, CA
reply to shdesigns
said by shdesigns:

I designed and built this in high school, probably 1977 or so. Made mostly from salvaged parts.

0-25V, 3 amps.

Why not keep the case and replace the insides for something like
0-25V, 0-10A?

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1
reply to mackey
said by mackey:

Hop on over to AliExpress and you can pick up a 20-pack of LM2596 chips for $5.80 shipped.

I know several people who use various inexpensive regulators based on LM2596 variations via Amazon. I could not find any obvious indication that their LM parts were counterfeit but they have original National Semi marking rather than new TI marking. Possibly old stock or recycled from old boards.

The DROK boards seem to be working fine in various applications for many months now. »www.amazon.com/DROK-Converter-Re···0JUFJ1GA

*** Anyway, Jameco's single unit price is only $4.99 for a genuine part: »www.jameco.com/1/1/23796-lm2596s···tor.html

LM2596-ADJ
Output Voltage 1.2V to 37V
Maximum Output Current 3A
Switching Frequency 173KHz
Switching Regulator Yes
Minimum Operating Temperature -40°C
Maximum Operating Temperature 125°C


shdesigns
Powered By Infinite Improbabilty Drive
Premium
join:2000-12-01
Stone Mountain, GA
Reviews:
·EarthLink
·Comcast
·Atlantic Nexus
reply to public
said by public:

Why not keep the case and replace the insides for something like
0-25V, 0-10A?

The case has lots of holes and is real cheap thin metal. I already have a 0-30V, 20A in a nice instrument case.

said by lutful:

but they have original National Semi marking rather than new TI marking. Possibly old stock or recycled from old boards.

Yes, these are marked with a NS logo. I don't see a date code on them.
--
Scott Henion

Embedded Systems Consultant,
SHDesigns home - DIY Welder


aurgathor

join:2002-12-01
Lynnwood, WA
kudos:1
Reviews:
·Frontier Communi..
reply to mackey
said by mackey:

And perhaps all of the above

Hmm, looking at this on eBay: »www.ebay.com/itm/like/201126191001?lpid=82

It's only $0.99 with shipping included!!!  
--
.sig


mackey
Premium
join:2007-08-20
kudos:13

1 recommendation

reply to lutful
said by lutful:

I could not find any obvious indication that their LM parts were counterfeit but they have original National Semi marking rather than new TI marking. Possibly old stock or recycled from old boards.

I did some Googling and found a thread over at eevblog.com where they discuss these chips; it seems there are a lot of counterfeits which use a 52-57kHz switching frequency (instead of 150kHz) and as a result have 200mV+ ripple.

I also remember a thread over in the WISP forum where someone bought like 10 of the cheap-o boards and all died after a few months, some taking the attached equipment with them.

/M

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1
said by mackey:

I also remember a thread over in the WISP forum where someone bought like 10 of the cheap-o boards and all died after a few months, some taking the attached equipment with them.

I was told that particular "someone" purchased 10 boards after that particular thread was already in progress. So I sort of doubt that claim.

Anyway, lower frequency operation can be easily checked.


shdesigns
Powered By Infinite Improbabilty Drive
Premium
join:2000-12-01
Stone Mountain, GA
Reviews:
·EarthLink
·Comcast
·Atlantic Nexus
reply to mackey
said by mackey:

I did some Googling and found a thread over at eevblog.com where they discuss these chips; it seems there are a lot of counterfeits which use a 52-57kHz switching frequency (instead of 150kHz) and as a result have 200mV+ ripple.

That's what this one does

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1

1 recommendation

For some reason, I could not find the actual EEVblog article that shows 52-57Khz operation. Anyone has the link?

Julian Ilett tested several modules. His reference power supply is also a Chinese module.

»www.youtube.com/watch?v=JLwJb4MVbls

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1
reply to shdesigns
said by shdesigns:

said by mackey:

it seems there are a lot of counterfeits which use a 52-57kHz switching frequency (instead of 150kHz) and as a result have 200mV+ ripple.

That's what this one does

Such low freq internal switching could be a rumour. Several expensive/genuine and cheap/fake LM2596 based adjustable boards appear to be using same value inductor.

Discontinuous mode operation may be responsible for the ripple - it is described in detail on page 22 of TI manual.


shdesigns
Powered By Infinite Improbabilty Drive
Premium
join:2000-12-01
Stone Mountain, GA
Reviews:
·EarthLink
·Comcast
·Atlantic Nexus

1 edit

1 recommendation

Click for full size
It is 55 khz. Image is 5us division

I measured 200mv ripple at 1.5A out.


mackey
Premium
join:2007-08-20
kudos:13
reply to lutful
said by lutful:

For some reason, I could not find the actual EEVblog article that shows 52-57Khz operation. Anyone has the link?

It's not a blog article, just some people discussing them in the forums.

/M

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1
reply to shdesigns
said by shdesigns:

It is 55 khz. Image is 5us division

I measured 200mv ripple at 1.5A out.

Can you kindly post a clear photo of your board? I am curious about the inductor and output capacitor.
You probably have a clone of this board shown in another of Julian's video.

»www.youtube.com/watch?v=6LTxJWsn1HQ


shdesigns
Powered By Infinite Improbabilty Drive
Premium
join:2000-12-01
Stone Mountain, GA
Reviews:
·EarthLink
·Comcast
·Atlantic Nexus
Click for full size
There seem to be two variants of this board on ebay.

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1
The apparent 1/3rd rate switching is quite puzzling. Assuming Chinese fab(s) copied LM2596 masks, the internal switching frequency should not deviate that much from 150Khz.

The inductor is only 33uH, same as many 150Khz applications, and should not work well at 1A+ output current or low voltage difference. But these boards are used widely in such applications.

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1
reply to shdesigns
said by shdesigns:

It is 55 khz. Image is 5us division

I measured 200mv ripple at 1.5A out.

Just wanted to post this since this thread could eventually become the original "proof" of apparent 55Khz switching of some Chinese LM2596.

Above formula is from appnote snva038b which also shows an example calculation. Estimate same 100 mOhm for their cap (it is probably much higher) and plug in 200mV ripple and 1.5A output current.

If internal switching frequency was really just 55Khz, that 33uH inductance is too low for many voltage/current settings where such boards are actually proven to be working.

Very puzzling.


mackey
Premium
join:2007-08-20
kudos:13
reply to lutful
said by lutful:

Assuming Chinese fab(s) copied LM2596 masks

As mentioned in the thread I linked to, they're probably actually LM2576's as those are much cheaper.

/M

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1

1 recommendation

said by mackey:

said by lutful:

Assuming Chinese fab(s) copied LM2596 masks

As mentioned in the thread I linked to, they're probably actually LM2576's as those are much cheaper.

I was just pointing out that even if they copied the masks and made zillions of clones in a local fab, the frequency is unlikely to have become 1/3 of original during the process. But if they are recycled or surplus stock, the internal "fixed frequency" oscillator will still be in 150Khz ballpark.

These PCB designers are copying TI appnote schematics for all these boards. But the inductor/ripple calculations do not match 55Khz operation for many proven working setups with higher voltage differential and higher output current. That is why I am puzzled.

public

join:2002-01-19
Santa Clara, CA
said by lutful:

That is why I am puzzled.

Nothing puzzling here. A Chinese vendor acquired a lot of reclaimed LM2576s, had them marked as lm2596, sold them to board stuffers selling ln2596 app note boards.

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1
said by public:

said by lutful:

That is why I am puzzled.

Nothing puzzling here. A Chinese vendor acquired a lot of reclaimed LM2576s, had them marked as lm2596, sold them to board stuffers selling ln2596 app note boards.

There is a major flaw with such innuendo ... an LM2576 switching at 52Khz could not use a small 33uH inductor to provide 1.5A output with just 200mV ripple. That is what shdesign reported when I was first puzzled.

I really do not want to argue this point with you, so here is the datasheet for LM2576:
»www.ti.com/lit/ds/symlink/lm2576.pdf


shdesigns
Powered By Infinite Improbabilty Drive
Premium
join:2000-12-01
Stone Mountain, GA
I will try to measure the inductor. It is marked 330, so it might just be 330uH. That is a reasonable value but the caps are too small.

lutful
... of ideas
Premium
join:2005-06-16
Ottawa, ON
kudos:1
said by shdesigns:

I will try to measure the inductor. It is marked 330, so it might just be 330uH. That is a reasonable value but the caps are too small.

It is a 33uH inductor. 33 x 10^0

A 330 uH inductor would be marked 331 for 33 x 10^1. Compact 330uH inductors that can handle 3A are very expensive.


shdesigns
Powered By Infinite Improbabilty Drive
Premium
join:2000-12-01
Stone Mountain, GA
I understand standard markings.

I have some inductors that are marked with their value instead of the nn X 10^n. It is more likely 33uh, and probably not real close