Hypex SMPS a minor Tweak

I have Hypex Ucd 400 amps with Hypex SMPS power supplies. They perform extremely well. Driving ATC SCM40 speakers and fed balanced from a Benchmark HDR DAC.
I was recently investigating the effect of using screened mains cable on parts of my system and noticed that when I used screened mains cable on my Hypex amps there was a slight improvement to the sonics.The cable used was 6A, 3 core mains + insulation +screen + outer insulation. Investigating further I concluded (correctly or incorrectly) that the Hypex SMPS supplies were outputting unwanted signals into the mains supply which were colouring the audio output of the amps.ie each was contaminating the other. I again concluded that the screened cable was acting as a form of filter and removing some of these unwanteds. I remade the mains cable,( 2 runs of app 4.5 M from each amp to the mains supply socket) using 10A screened instrumentation cable of the correct voltage rating. This cable was 3 core + screen + outer insulation so a higher capacitance /Metre than my earlier screened mains. I could not get a full tech spec on the cable, Capacitance was indicated at 100pf/M. So in theory a better filter.
The results were indeed noticeable a smoother treble and a little more detail, probably equivalent to about a 1.5 db improvement in sig/noise.
I have had another Hypex owner run totally independant tests with the same cable type and identical results were experienced.

I am now looking at a alternative to cable and experimenting with professional mains filters built into the IEC connectors I use as mains input connectors on my amps.
I have no experience of specifying these types of filters and the manufacturers data sheets do not show the filter response below 10Khz, where most of them are at 0db. The lower the current rating of the filter the better the bottom end response I presume due to the higher inductance value used. I am also assuming that the unwanted shash emanating from the Hypex SMPS is in the audio band + harmonics say up to 150Khz.
(I dont have a spectrum analyser)

Anyone give me advice on filter selection and am I progressing correctly

I am running fully balanced and my Hypex chassis is connected to mains earth.

The end result with my system was to identify the real culprit - the SMPS supplied with the Logitech Touch Streamer. I have now built a linear supply
which my DAC really appreciates and the end result is a well worthwhile improvement in my total system.

When I built my first Hypex I did email them regarding input filters and was told I didnt need them- anyway Ive learned a little more about mains contamination from both SMPS and Linear supplies.

Again can anyone keep my right on Filter Selection.

Ianmac
 
This is interesting.

I am planning a new "music centre" build. See my last efforts here: http://www.diyaudio.com/forums/digi...-music-centre.html?postid=1403097#post1403097

The new one will have a Transport > Tentlabs Clock > Monica 3 DAC > Nelson Pass B1 Buffer > 2x Hypex 180HG all in one box, served with one power cable.

The amps will be powered by the Hypex SMPS180.
The tranport (Marantz CD5000) will be powered by its original suppy.
The clock, DAC, and Buffer will be powered by a single 24vdc Meanwell SMPS, feeding each via individual Super Regs (Burson??? or similar) at appropriate voltages.

Ianmac: From your findings, do you think that there would there be excessive noise in the system from the Hypex SMPS? I do not have room for a linear supply, and I do not want external components.
 
No there would not be excessive noise. If you filter the mains in to the SMPS.it will improve things
Filters are readily available from Farnell or CPC farnell that are built into the IEC mains input socket. If you need details let me know.
I have e mailed Hypex re this issue and await their reply.
Ianmac
 
I have Hypex Ucd 400 amps with Hypex SMPS power supplies. They perform extremely well. Driving ATC SCM40 speakers and fed balanced from a Benchmark HDR DAC.
I was recently investigating the effect of using screened mains cable on parts of my system and noticed that when I used screened mains cable on my Hypex amps there was a slight improvement to the sonics.The cable used was 6A, 3 core mains + insulation +screen + outer insulation. Investigating further I concluded (correctly or incorrectly) that the Hypex SMPS supplies were outputting unwanted signals into the mains supply which were colouring the audio output of the amps.ie each was contaminating the other. I again concluded that the screened cable was acting as a form of filter and removing some of these unwanteds. I remade the mains cable,( 2 runs of app 4.5 M from each amp to the mains supply socket) using 10A screened instrumentation cable of the correct voltage rating. This cable was 3 core + screen + outer insulation so a higher capacitance /Metre than my earlier screened mains. I could not get a full tech spec on the cable, Capacitance was indicated at 100pf/M. So in theory a better filter.
The results were indeed noticeable a smoother treble and a little more detail, probably equivalent to about a 1.5 db improvement in sig/noise.
I have had another Hypex owner run totally independant tests with the same cable type and identical results were experienced.

I am now looking at a alternative to cable and experimenting with professional mains filters built into the IEC connectors I use as mains input connectors on my amps.
I have no experience of specifying these types of filters and the manufacturers data sheets do not show the filter response below 10Khz, where most of them are at 0db. The lower the current rating of the filter the better the bottom end response I presume due to the higher inductance value used. I am also assuming that the unwanted shash emanating from the Hypex SMPS is in the audio band + harmonics say up to 150Khz.
(I dont have a spectrum analyser)

Anyone give me advice on filter selection and am I progressing correctly

I am running fully balanced and my Hypex chassis is connected to mains earth.

The end result with my system was to identify the real culprit - the SMPS supplied with the Logitech Touch Streamer. I have now built a linear supply
which my DAC really appreciates and the end result is a well worthwhile improvement in my total system.

When I built my first Hypex I did email them regarding input filters and was told I didnt need them- anyway Ive learned a little more about mains contamination from both SMPS and Linear supplies.

Again can anyone keep my right on Filter Selection.

Ianmac

Hi Ian,

Can you post us some pictures from your actual setup? We have designed the SMPS400 to have as low as possible EMI, so I'm really surprised you have this kind of difference with or without a screened cable.

Regards,

Jan-Peter
 
Jan-Peter

I will be pleased to provide what you want.
It will be much easier for me to send pictures over to your address at Hypex. If you dont mind I will do that.
I have built 3 pairs of Hypex Ucd /SMPS monoblocs. 2 are here in use on my two systems. The third is in London being used by an audio enthusiast who carried out the independant test
All 3 systems perform well but the above tweak does make a noticeable improvement

Ianmac
 

barrows

Member
2010-07-23 11:28 pm
All...

The Hypex SMPS have AC filtering built in at the AC input. I suspect that this is the reason they generally recommend against adding additional power conditioning/filtering. The problem is, without measuring what one is doing when adding additional filtering, more capacitance across the line may interact with the capacitance/inductance of the built in filtering in a bad way (resonance). Without careful measurements, adding additional line filtering may do more harm than good. Longer distance of wire (R) between the filter elements (that is the built in filters in the SMPS and the added mains filter) may, in some cases be enough to prevent resonances, but without measuring, how can one be sure...
 
yes, these are hardly what I would call independent 'tests' where is the test?

there has been a lot of talk about SMPS spilling noise back onto the line, one of those fashionable critiques that seems to have caught on. i'm sure some do, but i'll bet there is a good chance the source of this problem lies inside the cases...

if the OP can notice 1.5dB difference in SNR on amps of this low SNR i'll eat my hat
 
Last edited:
Willem
My post was almost 2 years ago
Hypex were extremely helpful. There are minor differences between the early SMPS boards and the later ones circa Nov 2010. ( now doubt there will be differences between then and now) The detail of these difference I do not have but the later ones had a tab on the corner of the SMPS board that should be connected to ground, similar to the current Ncore SMPS supply.
I with assistance from a friend who had access to test equipment beyond my TeK scope, went on to look at the "filter effect " , as we saw, it of the screened cable. Cutting a long story short, I selected a filter I thought had the correct response and spoke to the manufacturer. These were obtained and built into a user friendly DIY assembly that allowed easy conection to equipment. Via another forum these were tested by quite a number of people on a variety of different source equipment with SMPS and Linear supplies in a variety of different mains connection set ups, mains supplies, polluted with or without unwanted signals. Hundreds of these have been built DIY.
My own experience, finally using my own best test equipment-- my ears-- was that my Hypex modules fitted with IEC input connector filters -- with my mains supply and my mains wiring gave me the best sound.
I used Schaffner FN9244-1-06 on my source gear and FN9244-3-06 on my Hypex
I hope that fills in the gap Willem

It would appear that my trying to quantify what I hear in sound quality improvement by using my humble experience in "probably equivalent to (x) db's" does not go down well on this forum with some wanting to eat hats. Perhaps I should follow the Hi fi pundits and use extravagent adjectives but then these havent been tested.

There is theory and there is practice.
The subject of noise emanating from power supplies both Linear and SMPS and possibly having an effect on adjacent equipment is well documented. Add to it the possible deficiencies of local mains supplies, different voltages (e.g. USA-UK) and equipment current consumption and pollution from a multitude of sources and one has a right good old mixture that will keep the theory boys busy for years - with indepenent tests - that might not get approved on this forum.

My own personal approach is that I now fit filters to everything I build. Sorry If I'm wrong and I dont have the test data to hand for every item of equipment.

I have not noticed any effects of resonance with my Hypex 4 hypex units, all of my linear supplies have an anti resonance network across the transformer secondary. My Hypex SMPS are as they left the factory.

Good filters cost only a few pounds. No one is forcing you to fit them. If they introduce other problems remove them. A standard IEC will fit into the hole left by a Filtered socket, not vice versa in most cases.