EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: HoracioDos on October 26, 2021, 01:50:05 pm
-
Hello.
I need to power a ICOM7100 and I can't find a linear power supply where I live, I only find cheap switching power supply from China and I don't want to spend time trying to filter the noise. So I need to build a linear one. Does anybody have a proven and simple design? I just want to build one and go.
Thanks in advanced
-
25-30A linear is going to be very big. You need a transformer that can handle that kind of current also very large capacitor.
-
Oldie but a good one, look for PW Marchwell or take a look at the Astron schematics for ideas.
-
Is this peak or continuous current? If you want cheap, hacked-together supply, you might be able to use car battery and a linear charger attached to it. Both are widely available. Lead battery will provide huge peak currents and good ripple filtering. 5 or even 10A linear chargers are available. There is also a benefit, that radio can be used for up to a few hours under blackout.
-
Hello Manul
25A continuous current @100W. I'm sure I will not have to worry about noise. LOL. If I get a 12V 30A battery and a floating charger I could start from there. As I'm getting my license I only do RX, So I only need 5A. I must connect RF ground as usual. Thanks!
-
69W for RX? That seems like a lot.
And what would be typical duty cycle when both TX and RX? 10A charger might be enough to keep battery from draining down in practical usage scenario, the only way to know for sure is to measure real world average current.
-
Are those those cheap surplus 12V server supplies quiet enough? Some are easily modified to output 13-14V, power tends to fall in the 750 to 1500 watt range.
-
Hello.
I'm listening WFM radio now and transmitter is taking 0.625A
-
Have a look at the Astron linear supplies. Very reliable, easy to repair and relatively inexpensive used. I have three RS-50A supplies. One powers a IC-7300, FT-8800 and Flex-6600 simultaneously (Only one transmitting at a time).
The only issue is high standby current and lower efficiency than switching supplies at low currents.
-
If you're happy to build it yourself, consider the VK Powermaster. Example - http://www.philipstorr.id.au/radio/seven/vkpower/VK%20Powermaster%2084.pdf (http://www.philipstorr.id.au/radio/seven/vkpower/VK%20Powermaster%2084.pdf)
-
A warning if you go to build something - older designs with 2N3055 transistors as the pass elements can often be unstable with modern transistors. This is easy to fix if it happens, just something to check for. It's because the 2N3055 specification is very broad with no upper limit on Ft, so much faster transistors made on modern processes can still meet the specification of and be sold as 2N3055.
-
Here is 13.8 V / 50 Amps PSU:
(https://i.imgur.com/NH0LqrQ.gif)
Tr1 should have 20 V no-load voltage and 16 V under 50 Amps load (after rectifier and capacitor).
Tr2 should have 24 V 0.5 Amps (after rectifier and capacitor).
here is description in Russian: http://www.ra4a.ru/publ/bloki_pitanija_i_zu/1/12-1-0-838 (http://www.ra4a.ru/publ/bloki_pitanija_i_zu/1/12-1-0-838)
Be careful with high currents, it can make fire and burns!
-
Wouldn't be that hard to rip the HV winding from a microwave oven transformer and add your own windings to give 13.8V at 30A.
They normally do over 1kW, for you 13.8V @ 30A is only 414W.
Granted, they don't normally run continuous and sometimes have a fan on them, but you only need 400W.
There's lots of DIY guides for turning microwave oven transformers into spot welders at ~2V or so
Just need ~7x as many turns.
-
Microwave oven transformers are a bit short on primary turns so run at very high flux densities, especially at light loads (which in a microwave is fine because they never have light loads). I don't think I would use one in this application, it'll run pretty hot even off load.
-
I almost entirely use Meanwell PSUs, high power linear power supplies are a pita; it's not worth bothering with them
in my Shack for my TS-2000X I use this:
https://www.mouser.ch/ProductDetail/MEAN-WELL/RSP-320-135?qs=%2Fha2pyFadujFkvpksmE%2F%252Bx4juEINt6RZLAjLGiMQXJJ4RgS82SPwTQ%3D%3D (https://www.mouser.ch/ProductDetail/MEAN-WELL/RSP-320-135?qs=%2Fha2pyFadujFkvpksmE%2F%252Bx4juEINt6RZLAjLGiMQXJJ4RgS82SPwTQ%3D%3D)
even if it's not declared as 25A continuous, it never let me down on my activities; and it delivers a very clean voltage; so no need for external filtering.
-
I almost entirely use Meanwell PSUs, high power linear power supplies are a pita; it's not worth bothering with them
in my Shack for my TS-2000X I use this:
Many ham amateurs using Meanwell PSU for transmitters in my location. It's easy to detect them by high background noise when they TX :)
It works, but it is very unpleasant to hear that noise.
-
If you're going to build one I quite like this design with the pass transistor collector common with the negative output. It gives you low dropout without an auxiliary transformer and also eases the heatsinking.
https://ludens.cl/Electron/Ps20/Ps20.html (https://ludens.cl/Electron/Ps20/Ps20.html)
Manfred also has a low-noise switching design for the ambitious builder:
https://ludens.cl/Electron/PS40/PS40.html (https://ludens.cl/Electron/PS40/PS40.html)
-
I have a 40A meanwell psu running the 14 meters of LED strip around the ceiling of the bedroom.
If the lights are on the FM radio is jammed all to hell.
-
Many ham amateurs using Meanwell PSU for transmitters in my location. It's easy to detect them by high background noise when they TX :)
It works, but it is very unpleasant to hear that noise.
do you have some hard facts for that? I know that Meanwell is often first choice in the ham domain, also by some very picky operators, who are fond of and engaged for their clean signals; so it'd surprise me
-
do you have some hard facts for that?
Yes, I hear this noise with my own ears. One time one of these amateur with Meanwell PSU got linear one for a short period of time. And when he tried to work with linear PSU, I notice very well that background noise disappears. The test was very impressive because the difference was just amazing. The sound of his station with linear PSU became crystal clear. With Meanwell PSU it was unpleasantly noisy.
Also he reported that he got much less noise floor with linear PSU and can listen much more weak stations, which is under the noise with Meanwell PSU. Before that case he advertised Meanwell PSU for TRX, but after that he is not so happy with it anymore :)
I know that Meanwell is often first choice in the ham domain
Yes, exactly. As I said above, many ham amateurs in my location using Meanwell PSU. But it doesn't means that this is better choice. This is just a cheap, lightweight and compact compromise. But it affects your receiver sensitivity and add noise on TX.
Some time ago I also tried to use Chinese SMPS for my QRP TRX. Cheap, small and nice... But later when I compared it with linear one, I never use SMPS for radio anymore. Because SMPS produce a lot of noise.
-
I almost entirely use Meanwell PSUs, high power linear power supplies are a pita; it's not worth bothering with them
in my Shack for my TS-2000X I use this:
Many ham amateurs using Meanwell PSU for transmitters in my location. It's easy to detect them by high background noise when they TX :)
It works, but it is very unpleasant to hear that noise.
I proved this to a couple of hams at my club by monitoring the TX fundamental of a 6M transverter with a spectrum analyser and switching between one such "quality" SMPSU and linear PSU, the noise was obvious and quite unpleasant.
I don't doubt it could be improved and filtered but it would take a lot of effort.
-
I also recommend checking the schematics for the various Astron linear power supplies which show how it is done with 2N3055s power transistors and a 723 regulator.
I have also done it using a PNP power transistor boosted LM317 with the board mounted directly to the large electrolytic input capacitors. The example below shows how this is done and is better than what I did at the time.
In either case, plan on using more than one power transistor in parallel to get sufficient power dissipation capability.
-
To me having a smps in the shack is like having dracula in charge of the bloodbank,dont go there!!!
-
To me having a smps in the shack is like having dracula in charge of the bloodbank,dont go there!!!
A switching power supply may be quiet enough to power radio equipment, however it is difficult to know ahead of time and without testing.
-
it depends on equipment sensitivity. But anyway good SMPS with low noise requires expensive filtering and shielding measures, so they are more complicated and much more expensive than linear one. So, it can make sense if there is very high power (for example 10-100 kW and above) and linear PSU will needs very expensive transformer.
-
I have tried multiple surplus SMPS, most of them are terrible and have all sorts of harmonics in the HF range. They are mostly built for cell tower stuff and the harmonics at HF freq are meaningless for this purpose. The SMPS I have that works well is a Ten Tec that is expensive. It is made for Ham HF equipment. I suppose that most of the SMPS built for Ham equipment would be better than normal SMPS. Eg: Yaesu, Kenwood, Icom etc....ALL SMPS produce noise, some of them just do not produce much noise in the RF spectrum you are dealing with.
Linear PS are usually much quieter and better, but harder and harder to find cheap. To build one you need a good big transformer to start, think several Kilos in weight. Then some big regulating transistors, and a good heat sink and fan.
I kinda like the battery solution, not only does the battery provide very clean power, It filters out much of the noise produced by the charger. So a cheap charger might be OK. Chargers can be very noisy since they are usually meant for short term use.
You can also run more than one battery in parallel for more power, you probably don't need to.
Gotta like "Dracula in charge of the blood bank"
Wally KC9INK
-
My thought is "it depends" and to be square, I'm not certain for all use cases. I like linear power supplies - especially if cost, weight, and size are not issues - but sometimes there are tradeoffs to be considered.
fwiw, I use Korad linear power supplies (KA3005 P and D) for "lab" work and they are spec'd and test out at around a 1mV of ripple. For radio work where more current is required I've used an Alinco DM-330MV (a switching power supply spec'd at 15mV for ripple) for lots of HF (mostly 20 meter and 40 meter) SSB and FT8 contacts as well as some VHF work, and I haven't detected an issue - but that's empirical vs tested experience.
I've also used a Mean Well RSP-1500-15 switching power supply which is rated at 150mV of ripple to make JT65 VHF (144 MHz) EME contacts (super weak signals with a roughly 250dB path loss between the Earth and Moon). I know another EMEer who has made a ton of EME contacts and he also uses a Mean Well switching power supply.
We are talking about signal noise, not fan noise here.
If we were talking about fan noise the Mean Well sounds like a small jet, the Korad makes some noise under a heavy load but is otherwise reasonably quiet, and the Alinco is pretty much audibly silent.
My power supplies were selected with consideration given to cost, size, and weight and of course to the current required (the Mean Well needs to support at least 50 and sometimes 70 Amps at about 13.8-14.1 VDC). If cost, size, and weight are not considerations and lowest noise is the goal it would seem that an Astron linear power supply would be a good choice. That's just a guess, I've never used an Astron.
I'd like to someday try to do some EME work with an Astron linear vs the Mean Well switching to see if that might help on the receive side - but my guess is the limitation is all the noise that surrounds the overall receive path of which the power supply is probably just a contributing component as the signal reaches the station. My preamp is integrated with the power amplifier which sits near the transceiver in the shack (rather than outside at the antenna), which requires about 50 feet of LMR600. Nonetheless, the Mean Well switching power supply has managed to support successful EME contacts with a 9 element yagi and ~300 watts on the transmit side, so how much of a big deal is the ripple? Like I say, I'm not completely sure. Might depend on the frequencies and signal strengths involved, the rest of the surrounding environment, and other variables in your particular use case.
-
We are talking about signal noise, not fan noise here.
Yes, we're talking about signal noise. With Meanwell SMPS noise is not very high, it don't override station signals. But it add background noise and it is very inconvenient to hear, especially for a strong signal which is expected to be crystal clear, but with Meanwell SMPS it has specific SMPS background noise (like buzzer with changing tone). Also it increase noise floor for a little.
-
We are talking about signal noise, not fan noise here.
Yes, we're talking about signal noise. With Meanwell SMPS noise is not very high, it don't override station signals. But it add background noise and it is very inconvenient to hear, especially for a strong signal which is expected to be crystal clear, but with Meanwell SMPS it has specific SMPS background noise (like buzzer with changing tone). Also it increase noise floor for a little.
So you are saying the noise is at or around some particular frequency? In the audible range of ~20Hz-20kHz, or at some frequency closer to a carrier frequency?
-
I wouldn't consider an SMPS for a decent radio - I experimented with one and a linear one on my IC-7300. The SMPS produced small regular peaks (spurs) on the spectrum display, the linear one was totally clean.
Some switched mode PSU's now have a knob so you can supposedly vary the switching frequency to minimise interference on the radio frequency you are working, but that just says to me "SMPS's are noisy here's a way to make it less bad but you need to fiddle with it when you change frequency or band"
I use a big(ger) heavy(er) linear but so what? It stays put, it's out of the way, it uses a few watts more, but it's QUIET
-
So you are saying the noise is at or around some particular frequency?
No, this is not single frequency. This is noise, but not white noise, some frequency components are predominant and their frequencies are constantly floating
-
So you are saying the noise is at or around some particular frequency?
No, this is not single frequency. This is noise, but not white noise, some frequency components are predominant and their frequencies are constantly floating
Yes Indeed, You can see the spikes walking right through the bands on many SMPS, if you just put an antenna on a spectrum analyzer near the SMPS, or on a radio with a spectrum display..
-
So you are saying the noise is at or around some particular frequency?
No, this is not single frequency. This is noise, but not white noise, some frequency components are predominant and their frequencies are constantly floating
Yes Indeed, You can see the spikes walking right through the bands on many SMPS, if you just put an antenna on a spectrum analyzer near the SMPS, or on a radio with a spectrum display..
I have an antenna on a spectrum analyzer. How close to the SMPS do you think is required to see the spikes? What center frequencies? How much span and resolution bandwidth? Thx
-
There is something else to consider if you are going to do your own design. Failure can be expensive if high voltage gets applied to your equipment, so include a crowbar circuit which shorts the output and blows the fuse if the output voltage rises above like 14.5 volts. The Astron linear power supplies mentioned earlier include crowbar circuit protection.
-
I actually used a HP Service monitor which has a sensitive receiver and a SA. I connected it to both my Ham antenna outside the house and one inside the house and a small UHF antenna that should be a bad antenna for 20 meters, which is where most of the crap was. The crappy 20 cm antenna did pick up all sort of spikes
On a SA I would just put a meter long wire on the input in the same room to start. Then move it closer.
I think most of my interference was on 20 meters .
Yea some of the linear PS have better protection than others
Wally KC9INK
-
To me having a smps in the shack is like having dracula in charge of the bloodbank,dont go there!!!
Powering the transverter from a converted server PSU gave results which were indistinguishable from the linear PSU but it's not a pretty solution and it's by no means guaranteed to be noise free, however they are a *lot* better than the cheap SMPSUs sold by a lot of Amateur radio suppliers.
It's worth knowing that a lot of commercial radio gear is and has been powered from switchers for a very long time, but, again, it's not the cheap nasty crap a lot of amateur radio suppliers punt.
-
It's worth knowing that a lot of commercial radio gear is and has been powered from switchers for a very long time, but, again, it's not the cheap nasty crap a lot of amateur radio suppliers punt.
As has most good quality testgear since at least the early 1990s, including equipment much more sensitive than a HF radio. Switchers can be done very well, but unfortunately often aren't.
it depends on equipment sensitivity. But anyway good SMPS with low noise requires expensive filtering and shielding measures, so they are more complicated and much more expensive than linear one. So, it can make sense if there is very high power (for example 10-100 kW and above) and linear PSU will needs very expensive transformer.
I would argue the crossover power is much lower than that, perhaps around 500 W. Yes filtering and shielding can be expensive, but so can thermal management, big pass transistors, and 50 Hz transformers.
-
Some switched mode PSU's now have a knob so you can supposedly vary the switching frequency to minimise interference on the radio frequency you are working, but that just says to me "SMPS's are noisy here's a way to make it less bad but you need to fiddle with it when you change frequency or band"
There have been PSUs fitted with them 'for radio use' for at least a decade, it's an indicator that screams cheap sh**e.
It is, as you say, an admission that the PSU it's fitted to is crap but it's not an indicator that all SMPSUs are.
It's worth knowing that a lot of commercial radio gear is and has been powered from switchers for a very long time, but, again, it's not the cheap nasty crap a lot of amateur radio suppliers punt.
As has most good quality testgear since at least the early 1990s, including equipment much more sensitive than a HF radio. Switchers can be done very well, but unfortunately often aren't.
Pretty much my point, yes, switchers can be really good, but to make them really good means you have to engineer them rather than just farting around with an app note or ancient design.
-
...but to make them really good means you have to engineer them rather than just farting around with an app note or ancient design.
And you would only engineer them to be low enough noise for radio use if their intended use was for something similarly sensitive. If some surplus server supply is particularly quiet that's cool, but that's probably mostly luck, or that the EMC test setup had a load of cable on the output so it had to be that quiet to pass. There should be good, low-noise PSUs on the market suitable for radio use, because it's technically feasible to make them at a similar or lower cost than a linear, at least at the higher-power end. But there's clearly a lot of crap out there too, and perhaps a reluctance to invest in developing a quality design for a market where the customers are quite anti-SMPS.
There is of course a separate issue of the quality of the radio gear. If it is sensitive to microvolts of noise on its power supply terminals and can only run properly on the most perfect of supplies, then arguably it is not a very good design. Any radio designer who assumes a perfectly noiseless 13.6 V d.c. input is clearly part of the problem - such equipment would presumably perform badly in a car with the engine running for instance.
-
The surplus switching power supplies (a few) that I have tried to use have both line noise and radiating noise. They make RF that is heard over the air right from the box as well as on the mains output.
I cannot remember the correct terms but engineers typically identify this noise as two types 1. noise on the power line and 2. noise coming over the air from inside the box itself.
Both of those problems need to be addressed in the engineering of the SMPS.
One very expensive surplus PS had a small SMPS board in it for "housekeeping" i.e. it powered the other circuits rather than putting out main power.
That "housekeeping" SMPS I believe was the source of the interference. You could put a low sensitivity receiving loop next to it and see the RF crap. Not so much near the other components.
The SMPS that I use is a Ten Tec that is made for HF radio and I do not see the interfering RF coming from it.
It is much more expensive.
A army engineer once told me that "ALL" SMPS throw RF, the challenge is to keep the RF out of the bands that the radio uses.
I think this is a mostly true statement.
Most of the surplus SMPS are designed for use in the UHF or higher bands (cell phone stuff) or for use with computer equipment. Interference in our HF HAM bands would not be critical to the successful design.
-
I actually used a HP Service monitor which has a sensitive receiver and a SA. I connected it to both my Ham antenna outside the house and one inside the house and a small UHF antenna that should be a bad antenna for 20 meters, which is where most of the crap was. The crappy 20 cm antenna did pick up all sort of spikes
On a SA I would just put a meter long wire on the input in the same room to start. Then move it closer.
I think most of my interference was on 20 meters .
Yea some of the linear PS have better protection than others
Wally KC9INK
I ran a wire from the SA to within a few inches of the Mean Well power supply. There is a noise floor around -80 dBm that doesn't seem to change or show any spikes whether the power supply is on or off. I looked around 20 meters (14.2 MHz with a 4 MHz span). Maybe the Mean Well is a decent performer or maybe I'm not measuring properly.
-
Dont overcomplicate things!
For HAM use you need a fixte voltage of about 12-14V this can be done pretty simple.
First you need to get your hand on a transformer something that will give out about 18V-15V. This you can modify a micro transformer or canabalice an old 20/30A 12V battery charger.
The voltage regulating circuit is simplest to just have a 7812 regulator and some boost transistors.
This is an exsample of one way to build the PSU.
https://320volt.com/en/l7812-2n3005-ile-12v-15-volt-20-amper-guc-kaynagi/ (https://320volt.com/en/l7812-2n3005-ile-12v-15-volt-20-amper-guc-kaynagi/)
Here is another.
https://www.circuitstoday.com/12v-15a-voltage-regulator (https://www.circuitstoday.com/12v-15a-voltage-regulator)
The output voltage on these will be 0,7V under the output of the regulator.
JUst one last thing te reccomended load current for these applications is for a 2N3055 about 5A for each transistor, but you can just paralell them fo get more current.
The absolutely simplest way is to use a 317 regulator this willgive you the option of finetuning the output voltage.
https://electronicsarea.com/20-amp-variable-power-supply-using-lm317/ (https://electronicsarea.com/20-amp-variable-power-supply-using-lm317/)
-
Thanks for all for replies. I bought a 12v -10A battery yesterday and a floating charger. I'm still reading all your post very carefully. This design is very simple to build with a LM741 https://ludens.cl/Electron/Ps20/Ps20.html I would like to start from here and check results later.
Thank you all again!
-
That is very simple, and may suit your needs, but it really ought to have a few extra components to protect your ICOM in case one of the pass transistors should fail. Look up circuits for Crowbar protection; these will protect your rig if the voltage from the charger exceeds a set limit (e.g. 14.5V). I think it best to not depend on the battery to keep the volts down.
-
Thinking about this some more, using a 10Ah battery means you can only charge at 1 or 2 amps and so I think that battery will drop in voltage very soon when it starts seeing the Tx load. I recall running a 75W Tx on a 40Ah battery with a 5A charger many years ago and that just about worked, at least, for a substantial period of time. The Tx will take current from the battery rather than the charger because it's a lower internal resistance.
-
Hello StuartA
I know that a 10A battery could not be enough for TX, but as I'm getting my license I only do RX. Meanwhile I'll get or build a proper PS. I'm also interested in QRP so I think that this battery won't be a total waste. At least I can use it for my UPS. Thanks for your comments.
-
Point taken about Crowbar protection. Something like this
-
Point taken about Crowbar protection. Something like this
That is exactly right. The small ceramic/film capacitor prevents RF from triggering the SCR, and I think really should be at the gate, and the larger capacitor delays triggering slightly from expected ripple.
-
Crowbar circuits are good and all but if you have a PSU capable of supplying a constant 30 amps you need one hell of a thyristor and fuses don't blow instantly.
There are other ways to do overvoltage protection, this for example is 'in progress' here (waiting for boards and chips)
https://gm4wzg.co.uk/wp/home/projects/bob/
PW Marchwood PSU had an interesting scheme too, it used a crowbar to make a relay drop out but it left the reservoir caps connected so the overvoltage would still be present until they drained, however it could easily be adapted to disconnect the output instead.
https://worldradiohistory.com/UK/Practical/Wireless/80s/PW%201983%2006.pdf
-
Crowbar circuits are good and all but if you have a PSU capable of supplying a constant 30 amps you need one hell of a thyristor and fuses don't blow instantly.
It is just about an ideal application for a fuse since the peak current is so high, and thyristors have excellent peak current capability, and thyristors are inexpensive. The zener type trigger also applies good drive although this could be improved with some type of positive feedback switch, like another SCR?
-
Crowbar circuits are good and all but if you have a PSU capable of supplying a constant 30 amps you need one hell of a thyristor and fuses don't blow instantly.
It is just about an ideal application for a fuse since the peak current is so high, and thyristors have excellent peak current capability, and thyristors are inexpensive. The zener type trigger also applies good drive although this could be improved with some type of positive feedback switch, like another SCR?
It's an excellent application for a fuse, yes, they're designed for the sort of current handling profile needed.
Which is why they don't blow instantly.
I'm not suggesting crowbars are redundant, I'm suggesting there are other ways to achieve results.
I'm a fan of a crowbar circuit as a last resort but I've had to replace shorted thyristors and repair burned up boards more than a few times when a PSU has gone OV.
-
Point taken about Crowbar protection. Something like this
That needs to have a fuse between the transistors and the zener diode. The idea is that if the voltage goes too high, the thyristor conducts, bringing the output voltage to zero, but in so doing, blows the fuse. It may take some milliseconds to blow the fuse, but that is fast enough to mean that the thyristor should only need a very modest heat sink. However, I wouldn't use a PCB for a Crowbar circuit, which really needs some hefty wires.
-
That needs to have a fuse between the transistors and the zener diode. The idea is that if the voltage goes too high, the thyristor conducts, bringing the output voltage to zero, but in so doing, blows the fuse. It may take some milliseconds to blow the fuse, but that is fast enough to mean that the thyristor should only need a very modest heat sink. However, I wouldn't use a PCB for a Crowbar circuit, which really needs some hefty wires.
With a printed circuit board, a constriction in the high current trace will cause it to fail there first if the fuse does not blow quickly enough.
-
A few years ago i put together a linear psu for my radios,it ran my ts-850s and tr-9000 fine,but i got an old icom fc-1010 ex taxi radio and reprogramed it on 2mtrs,was funny it worked fine on a car battery but had about a 2khz sqeal when the squelch closed when running on my homebrew psu,it was a lm723 design,guess it must of been oscillating?