• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Switched Mode Power Supplies (SMPS)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Thomas,
please have a try first with the existing capacitors when the boards arrive. if the overcurrent triggers, let me know and i will tell you what how to deal with.

Esperado,
please wait till next monday, i will try to clarify the product description for each power supply.


Because of some recent unpleasant experiences i have to change the policy regarding the custom available voltages for power supplies. There will be a list with clear defined available voltage values and will be specified with BOLD characters the voltage values on request will require extra time and extra cost. that's because there are just few ppls who really appreciate this and a lot more who only can complain and send mails after mails, one-two-three days after the ordered a power supply asking why it wasn't sent yet, without care that their transformer will take another week or two and 10-20 usd more to be done. I'm not getting rich out of this, and i don't need extra headache and complains every time someone expect his product to be shipped before can actually be done. then there are shipping problems, recent Universiade games in Shenzhen, which brought mess and delay. some packages were stuck at customs for few weeks, without being able to be sent out. it is out of my control what happens after i post a package, and one should understand that i cannot send every package with DHL or UPS. otherwise for some products the shipping will cost more than the product itself.
please read this before order something. and mail me for any question. thank you.
 
please read this before order something. and mail me for any question. thank you.
Thanks, Cristy.
Don't worry, and don"t care with unreasonable customers. I don't believe, if some order any tension custom tension, he don't understand the transformer has to be made, and the SMPS tuned !
But i believe, for sure, that you are both an experienced electronician and a very very nice guy.
My remark was just to help you, then your customers, to find easier the right product from you, and, so, increase your business ;-)
May-be it would be good to separate DC-DC converters from AC-DC SMPS. In the page where they are all listed, to add a line like:
+&-Xv toXv, +yV -yV under the reference.
And a table with all the available tensions, currents and noise for each PSU in the product page.

Well, because i do not used any SMPS for audio, i asked this question: Can-i power on the same SMPS a Class D amp (~300khz) and a linear Class AB one with a emitter follower extra filtering (for treble) without to worry about the 300khz ripple, or did i better use two separate PSUs ?
And an other question about Auxilliary. May i add an other rectifier/regulator branch in order to get -12 V ? In other words, is the auxiliary current limited mainly by the transformer ? And what is the current limit ?

Thanks for your care.

What is the noise measurement on this?
Someone had measured-it on a French forum to -70 to -90db around 125Khz and 10Mhz
Very impressive. It needed to put the SMPS board in a Faraday cage, the noise is mostly RFI.
 
Last edited:
SMPS800R arrived!

My 2 x smsp800r arrived today:)!

This might be a silly question but how do I power them up. I will have an external power switch on my case so I simply bridged the switch connection.

When I connect 120V nothing seems to happen on either board. What am I missing:confused:

The IEC connectors were not soldered in at my request and and I removed them to solder the supply wires directly onto the board.

Cheers

Thomas
 
Thomas,
the 110V jumpers are missing, you can see the place for them just in front of the big mains side capacitors. they must be connected together for 110V operation.
recently, i do not connect them, and i let the user connect himself, if he is sure that the available mains voltage is 110V. that's because there are countries with 110/220V like Brazil or the user might live in two countries, like Canada and France.
 
Thomas,
please have a try first with the existing capacitors when the boards arrive. if the overcurrent triggers, let me know and i will tell you what how to deal with.

Esperado,
please wait till next monday, i will try to clarify the product description for each power supply.


Because of some recent unpleasant experiences i have to change the policy regarding the custom available voltages for power supplies. There will be a list with clear defined available voltage values and will be specified with BOLD characters the voltage values on request will require extra time and extra cost. that's because there are just few ppls who really appreciate this and a lot more who only can complain and send mails after mails, one-two-three days after the ordered a power supply asking why it wasn't sent yet, without care that their transformer will take another week or two and 10-20 usd more to be done. I'm not getting rich out of this, and i don't need extra headache and complains every time someone expect his product to be shipped before can actually be done. then there are shipping problems, recent Universiade games in Shenzhen, which brought mess and delay. some packages were stuck at customs for few weeks, without being able to be sent out. it is out of my control what happens after i post a package, and one should understand that i cannot send every package with DHL or UPS. otherwise for some products the shipping will cost more than the product itself.
please read this before order something. and mail me for any question. thank you.

On your site:

Since you ship worldwide why not say on your site shipping to US takes 3 weeks (or whatever) to Brazil 2 weeks (or whatever).

Since you don't respond to emails on your site very often, why not say so?

If you ship products only once a week or every 2 weeks why not say so?

Put some connection info for 110v and 220v, and say you don't solder the IEC connector.

Adding grounding info would be good too.


More info is better on your site. Customer expectations will then match what you do.
 
Thank you for the very useful hints. let me complete some.
-shipping time vary from country to country and i don't remember when the shipping company told me last time that will take more than 2 weeks to any destination. and if there is no special event or public holiday, it does not delay. Right now are some packages which haven't left china and they are moving from Shenzhen to Xiamen then to Shanghai, back to Guangzhou and eventually sent out. HK post messed up few times, they lost some packages and i decided to not use them since then. DHL and UPS are alternatives for expensive or urgent shipping. but they cost too much to afford to send each package with them. everytime when there were some shipping problems i changed the shipping company. 6-7 times in the past 3 years. i realized that none is better than the previous one.
-i reply to mails every day, except days when i;m not in Shenzhen. i spend 2-3 hours doing this. i get 40,50, 60 mails a day, excluding spam. have you sent me a mail and i haven't reply ? sent it again. is free.
i ship products 2-3 times every week. in exceptional cases once a week, but NEVER at two weeks, even during chinese new year i had to go to HK to send packages.
-90% of the customers asked me to NOT solder the connector. first batch of boards had the connectors ready soldered and i had to desolder them all. wasn't funny.
-i will add more infos when i will be able to access the site. imagine that you (me) live in a world where internet is a luxury, although i have it at every step.
 
SMPS800R for Class A/B amp

I finished one channel of the AMB β24 with the SMPS800R. Slow start current limiter works and everything seems to be fine.

Replacing the single big toroid and filter caps with individual SMPS800R per channel saves about 3kg from the build and turns it into a dual mono layout.

Cheers

Thomas
 
Last edited:
I finished one channel of the AMB β24 with the SMPS800R. Slow start current limiter works and everything seems to be fine.

Replacing the single big toroid and filter caps with individual SMPS800R per channel saves about 3kg from the build and turns it into a dual mono layout.
Changes in the sound ? Did your dual mono had a negative effect against the stereophonic image ?
 
Here is a picture of the switching ripple from the SMPS800R. Peak-2-peak this is almost 4V. Any ideas for a filter to reduce by at least an order of magnitude, ideally less than 100mV.

Just for reference to output of the VAS stage power supply next to it.

Would this filter be worth trying out http://www.trcelectronics.com/Cosel/pdf/snr-10.pdf?

Cheers

Thomas
 

Attachments

  • P9056694.jpg
    P9056694.jpg
    191.2 KB · Views: 354
  • P9056696.JPG
    P9056696.JPG
    139.6 KB · Views: 326
Last edited:
Yes I am pretty sure. The two power supplies are mounted onto the same heatsink and share a common ground.

The first is picture is taken from the plus output of the SMPS800R

The second picture is taken from the VAS stage power supply which happens to be a Sigma 22 from AMB.

The settings on the scope are untouched.

Cheers

Thomas
 
Yes I am pretty sure.
Again, it is absolutely indispensable to enclose the SMPS. There was one thread on a french forum where somebody, as sure as you, had tried everithing to filter this SMPS. Then he had shielded-it, and the real noise was, if i remember well, around 30mv without any additional filtering. Better no additional filtering was able to decrease this noise level, often increased on the contrary. Your noise looks exactly like his one, before the grounding.
 
Okay, so I have common ground and the two power supplies are mounted directly next to each other. The measurements are taken an inch away from each other.

Not very clear how RFI at the order of 2V swing would impact one but not the other.

Coming up with a cage is not so hard, so maybe I try that. Ventilation is going to be an issue though. Do you have a link to that forum post?

Cheers

Thomas
 
Coming up with a cage is not so hard, so maybe I try that. Ventilation is going to be an issue though. Do you have a link to that forum post?
Projet : Ampli à base de LME49830 - "Gran Torino 150" - Page 49 - Amplification.
He had done it with a simple mesh metallic plate.
It was an amusing thread to read for me, because he was searching a power supply noise problem, when it was evident from the beginning this was an RFI problem on the input line of his amplifier, coming from the SMPS.
Note that he got, with twisted power wires between SMPS and Amplifier: 32mV of noise, not so bad ? JWithout twisted wires, it was 37.5mV .
 
Last edited:
The real ripple of any power supply can be measured properly in some certain conditions. the most important is the oscilloscope probes quality and the way how is connected to the measuring point, the output connector.
The probe must always use the shortest GND return path, without the clip cable and with the loop made of tip and shield to be not more than 10mm.
Here are some interesting articles http://www.eet-china.com/ARTICLES/2003AUG/PDF/2003AUG06_BT_ST_HBM_DA_AN05.PDF
How to Measure Power Supply Output Ripple
Power Tip #6: Accurately Measuring Power Supply Ripple
if the ripple was measured with the GND clip connected on the pcb, then the fringing flux of the transformer induced the noise which can be seen in loop formed by the proble GND cable. the topology used in this power supply, LLC resonant converter lead to strong fringing flux in the transformer, and this flux can be contained by a simple shielding of the board or installing with proper clearance to the small signal stage . i would not recommend to try to shield the transformer alone, this will change some important parameters as leakage inductance and would pose safety problems. the radiated EMI is in plan with perpendicular axis on the pcb and decrease exponentially with the distance.
 
The real ripple of any power supply can be measured properly in some certain conditions. the most important is the oscilloscope probes quality and the way how is connected to the measuring point, the output connector.
The probe must always use the shortest GND return path, without the clip cable and with the loop made of tip and shield to be not more than 10mm.
Here are some interesting articles http://www.eet-china.com/ARTICLES/2003AUG/PDF/2003AUG06_BT_ST_HBM_DA_AN05.PDF
How to Measure Power Supply Output Ripple
Power Tip #6: Accurately Measuring Power Supply Ripple
if the ripple was measured with the GND clip connected on the pcb, then the fringing flux of the transformer induced the noise which can be seen in loop formed by the proble GND cable. the topology used in this power supply, LLC resonant converter lead to strong fringing flux in the transformer, and this flux can be contained by a simple shielding of the board or installing with proper clearance to the small signal stage . i would not recommend to try to shield the transformer alone, this will change some important parameters as leakage inductance and would pose safety problems. the radiated EMI is in plan with perpendicular axis on the pcb and decrease exponentially with the distance.


Good info to know.

Can you change the SMPS800R output to +-65v?
How much time?
How much extra cost?

Thanks
 
The real ripple of any power supply can be measured properly in some certain conditions. the most important is the oscilloscope probes quality and the way how is connected to the measuring point, the output connector.
The probe must always use the shortest GND return path, without the clip cable and with the loop made of tip and shield to be not more than 10mm.

That is very interesting and I will go and measure directly at the connector with the ground clip off and a 10mm connection to the probe shield.

However for practical purposes in an amp, I will have to have a longer ground connection to the start ground. When I measure the voltage on the amplifier board the ripple is very much there.

In fact if I look at a 1Khz output signal at the speaker outputs the ripple is clearly visible in the waveform.

I guess I will have to try the cage.

Cheers

Thomas
 
I modified the probe from the scope to get the ground connection with a short piece of wire directly at the tip and connected that setup to the output terminals of the SMPS800R.

The result is indeed identical to what Gilles saw in the thread on the french forum but I am getting a little higher peak-to-peak riple of 270mV. The same procedure for the Sigma 22 gets about 20mV peak-to-peak swing.

While this is progress from the 4V yesterday I still don't understand how to connect wires to the device to reach the star ground and the power connector of the amplifer board without introducing a large amount of ripple into the amplifier board.

This leaves me with a question for Cristi. You suggested that enclosing the power supply in a Faraday cage might interfere with it's operation. On the other hand there seems to be evidence that this dramatically reduces the switching ripple, so is it safe to do it?

Cheers

Thomas
 

Attachments

  • P9066701 (800x600).jpg
    P9066701 (800x600).jpg
    138.6 KB · Views: 442
  • P9066698 (800x600) (750x563).jpg
    P9066698 (800x600) (750x563).jpg
    159.5 KB · Views: 438
Last edited:
This leaves me with a question for Cristi. You suggested that enclosing the power supply in a Faraday cage might interfere with it's operation. On the other hand there seems to be evidence that this dramatically reduces the switching ripple, so is it safe to do it?

Cheers
Thomas

Cristi said:
i would not recommend to try to shield the transformer alone, this will change some important parameters as leakage inductance and would pose safety problems. the radiated EMI is in plan with perpendicular axis on the pcb and decrease exponentially with the distance.

If I may, cristri said putting the onboard transformer alone in a faraday cage may cause problems. Putting the entire SMPS in a cage can not harm.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.