Received my PCBs last Friday, thanks! Now, all I need is some time to order components, to build regulators and to do some measurements.
Hi raoul,
Have you managed to build and test this PCBs?
BR,
Aleš
AofE third edition compares lots of regulators and gives some great information--for free in this case!
See chapter 9 pdf: http://artofelectronics.net/wp-content/uploads/2016/02/AoE3_chapter9.pdf
Build my power supply last week based on the LM338 regulator on Ales pcb.
The only parts which i didn't mount on the pcb were the smoothing capacitor and the bridge rectifier.
Build mine for 5V/5A and tested it with 50 ohm load (100mA), receiving these results between 0-100kHz:
Ripple 5,5µVrms
Noisefloor -105dB
PCB Parts:
http://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=9380d184f6
Bridge rectifier Vishay GBPC3502W
C1: Smoothing Capacitor 10000µF
LM338 regulator
C2: 0.1µF solid tantalum SMD 1206
C3: 10µF solid tantalum SMD 1206
C4: 10µF solid tantalum no SMD
R1: 120 ohm 0,1%
R2: 5k ohm trim 5%
D5/D6: 1N4007
2x Led's red
As you see in the parts table i only use very good components with the tightest tolerances to receive maximum performances. Also a very good medical EMI/RFI inlet is used in this power supply.
You need also a stable transformer (i used the 160VA Amplimo version).
The only parts which i didn't mount on the pcb were the smoothing capacitor and the bridge rectifier.
Build mine for 5V/5A and tested it with 50 ohm load (100mA), receiving these results between 0-100kHz:
Ripple 5,5µVrms
An externally hosted image should be here but it was not working when we last tested it.
Noisefloor -105dB
An externally hosted image should be here but it was not working when we last tested it.
PCB Parts:
http://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=9380d184f6
Bridge rectifier Vishay GBPC3502W
C1: Smoothing Capacitor 10000µF
LM338 regulator
C2: 0.1µF solid tantalum SMD 1206
C3: 10µF solid tantalum SMD 1206
C4: 10µF solid tantalum no SMD
R1: 120 ohm 0,1%
R2: 5k ohm trim 5%
D5/D6: 1N4007
2x Led's red
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
As you see in the parts table i only use very good components with the tightest tolerances to receive maximum performances. Also a very good medical EMI/RFI inlet is used in this power supply.
You need also a stable transformer (i used the 160VA Amplimo version).
I think tantalum capacitors are the WORST made of all. I have seen them fail 50x more often than any other type. As DIY guru Rod Elliott says:
"Then of course we have tantalum electrolytics. While many sing their praises, I do not recommend their use for anything, other than tossing in the (rubbish) bin. There might be the odd occasion where you really need the properties of tantalum based caps, but such needs should be few and far between. They are unreliable, and have a nasty habit of failing short-circuit. They cannot tolerate high impulse currents and/or rapid charge/ discharge cycles, and especially don't like being shorted. Tantalum caps announce their failure by becoming short-circuited, and it can be extremely difficult to track down a (possibly intermittent) short across a supply bus that powers many ICs. I never use tantalum caps, and don't recommend them in any of the published projects. Personally, I suggest that you don't use them either."
This is quite outdated information. Nowadays tantalum caps have good to excellent properties. I have seen many drop type tantalum caps gone to heaven with a loud bang and a nasty smell but this was at least 15 years ago, way before SMD existed. All the things Rodd says were true for those. I could never understand designers using them directly on power supply rails with their bad habit of shorting out. British audio was loaded with tantalum caps (also in the signal path that caused some cult following) which were the first suspects when something was wrong and indeed they proved to be unreliable. This gave them a very bad reputation. Every guy repairing stuff would toss them out at first sight as "if they don't fail now they definitely will fail later". Believe me, I would never have thought they would become a nice alternative but they are. It seems certain facts from long ago will linger on for ages in audio circles
With modern low ESR tantalum caps I have yet to experience the first one failing. There is practically no smart phone without tantalum caps as these are small with large values and excellent features. They can't dry out and can withstand high temperatures. I use AVX TPS series very frequently and can only recommend to try those out.
* Most tantalum caps don't accept abuse with regards to polarity but this also counts for electrolytic caps. Why manufacturers have chosen to label the + pole is beyond logic. Normal electrolytic capacitors have the - labelled. Many a DIYer has soldered or will solder them in reverse which will reconfirm the outdated and nowadays untrue reputation.
With modern low ESR tantalum caps I have yet to experience the first one failing. There is practically no smart phone without tantalum caps as these are small with large values and excellent features. They can't dry out and can withstand high temperatures. I use AVX TPS series very frequently and can only recommend to try those out.
* Most tantalum caps don't accept abuse with regards to polarity but this also counts for electrolytic caps. Why manufacturers have chosen to label the + pole is beyond logic. Normal electrolytic capacitors have the - labelled. Many a DIYer has soldered or will solder them in reverse which will reconfirm the outdated and nowadays untrue reputation.
Last edited:
This is quite outdated information. Nowadays tantalum caps have good to excellent properties. It seems certain facts from long ago will linger on for ages in audio circles
Perhaps you are right; but Mr. Elliott's article (Capacitor Characteristics) is marked as "Updated August 2013". So I don't know if it is "outdated"
I know it is. Please note that producing defective parts won't make a manufacturer last 30 years on the market.....
The industry also chooses tantalum caps for their properties despite their higher cost and all the bad stuff about obtaining coltan. Some switched to niobium which comes from the same areas. Niobium caps are also a nice alternative.
It is the same as saying american cars are junk. They once were and nowadays they still.....eh......
The industry also chooses tantalum caps for their properties despite their higher cost and all the bad stuff about obtaining coltan. Some switched to niobium which comes from the same areas. Niobium caps are also a nice alternative.
It is the same as saying american cars are junk. They once were and nowadays they still.....eh......
Last edited:
Hi Fortitudine
I am sorry I missed your post earlier.
Excellent reference indeed, thank you very much.
All, see page 618 for noise
BR
Jean-Louis
It's all in the LM338 specs which capacitors you need to use.
http://www.ti.com/lit/gpn/lm338
Don't forget the very good EMI/rfi filter inlet.
10.000µF is the lowest value you will need for proper working !
I can explain that in the following examples:
At 230V with no load
For 5V i need a 9V transformer.
Bridge rectifier = 1,1V
Usec=(1.41 x 9)-1.1 = 11,6V
Uc1=11,6 x 1,4 = 16,2V
I need 5V dc output, LM338 has 2,75V dropout voltage makes: 7,75V top
Max ripple at the LM338 input = 16,2 - 7,75 = 8,5V
Imax = 5A
C=Ic*0,01/Uc1
C=5 x 0,01/8,5 = 5900uF
This is without load !!!
At 230V with load
For 5V i need a 9V transformer.
Bridge rectifier = 2x1,1=2,2V (double rectifying)
Usec=(1.25 x 9)-2,2 = 9,05V
Uc1=9,05 x 1,4 = 12,67V
I need 5V dc output, LM338 has 2,75V dropout voltage makes: 7,75V top
Max ripple at the LM338 input = 12,67 - 7,75 = 5V
Imax = 5A
C=Ic*0,01/Uc1
C=5 x 0,01/5 = 10.000uF
So at full load (5A) you need at least 10.000uF capacitor !!!
There is no R3 ?
If you going to use the LM338 i would recommend to keep the values which are in the spec of the LM338.
If you going to use another regulator, for example the LT1084, the values are also different in the specs.
Best regards,
I think the ripple will be increased at higher voltages, but after all i'm very happy with the results now.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.