Silicon Chip Headphone Amp Tweaks

[johnmclean]

Thanks very much Alex, I didn’t t see your PS on the previous post, sorry I got you to repeat yourself. That all makes perfect sense now, hope to see you next week at Nicks.

Cheers
John

[PinkFloyd]

I don't know if this has already been tried but I thought I'd give it a go after Alex mentioned it in a seperate thread......

Fit a MOV varistor between live and neutral:



A definite improvement! lower noise floor, deeper bass, silkier mids, improved detail... I jest not, this little disc has made a very noticeable improvement.

Of course it makes sense, this little device absorbs high-voltage surges (and these can occur thousands of times in one day) so the voltage entering the amp remains more stable.

The MOV I have used is an EPCOS 275AC type (Energy Absorption: 151 Joules) these are about the size of a two pence coin and can be bought for around 50 pence so certainly won't break the bank.

I have a box of them and will be happy to supply these at part cost plus postage and packing, PM for details if you are interested.

[Deleted]

is that an improvement over and above the X2 cap?
Is that joule rating ideal or random?
Also what cap did you use with it?

[Deleted]

Chris
I can't answer the 1st question, but I think the main reason that Mike tried one here in place of the capacitor, was due to the occasional RIFA high voltage capacitors failing in the Little Pinkies.(faulty batch ?) Joule rating shouldn't matter much in this application, like it would where you are using them in mains filters or power boards for surge/over voltage protection.
Incidentally, I use these devices in a heavy duty mains filter that supplies all my AV gear.
If I had got around to usng this filter in line when I first moved to my current address, I would most likely not have had a couple of transformers damaged due to an over voltage event during construction of a nearby new power sub station.
Alex

[Deleted]

Alex,

Thanks for the reply;

I think the main reason that Mike tried one here in place of the capacitor, was due to the occasional RIFA high voltage capacitors failing in the Little Pinkies

Yes, I was aware of the reasoning but was curious if Mike found the varistor and cap to be sonically superior, in the scha and the little pinkies, to the X2 cap.

On opening the surge protectors I use at work I saw these varistors in there too. The hifi has UPSs protecting it.

[loupy31]

Hi All,
Two questions, first, Do I have to use heat sinks on the 5171 and 1930's which are on the JLH board?

second, The MOV Varistor, in the pic above is placed across the live and neutral terminals, The X2 cap is placed across the power switch, Basically one is in series and the other in parallel, are they both required or do I use just one, and if so which one.

Regards Peter

[Deleted]

Aug 16, 2009 9:36:45 GMT loupy31 said:

Hi All,
Two questions, first, Do I have to use heat sinks on the 5171 and 1930's which are on the JLH board?

second, The MOV Varistor, in the pic above is placed across the live and neutral terminals, The X2 cap is placed across the power switch, Basically one is in series and the other in parallel, are they both required or do I use just one, and if so which one.

Regards Peter

Peter
Unless you live in a place like the U.K. or Scotland , where our winter days are often warmer than their summer days ,
I would recommend that you fit the small heatsinks as supplied with the kit.
The X2 capacitor across the power switch is to help eliminate contact arcing, which MAY manifest itself as a very loud click when the switch is operated or turned off. The varistor has a different purpose in life, and greatly reduces the effects of high voltage spikes coming in on the mains supply. It also gives quite a bit of protection to the amplifier in the case of short term high overvoltage on the mains supply. Longer term , it should enable a properly rated fuse in the IEC socket to blow, thus hopefully protecting the equipment from major damage.
Alex

[loupy31]

Thanks for that Alex, What value should i use for the X2.

Peter

[Deleted]

Aug 16, 2009 10:02:45 GMT loupy31 said:

Thanks for that Alex, What value should i use for the X2.

Peter

Peter
Silicon Chip SUPPOSEDLY used a 1nF 250VAC X2 in the Stereo preamp that the original HA was part of, in conjunction with a 500mA fuse. However, I can not see it in the photo or wiring diagrams in the magazine article. Also, the lowest value X2 capacitors from Jaycar and Altronics web sites appear to be 10nF.
Alex

[loupy31]

Hi All,
Just another quick question. As far as AC supply goes, Do I need to use a 3 wire lead from the wall socket to the unit (IEC), which would include earthing to the case/enclosure, Or can it be a 2 wire Ac cable into a 2 pin AC socket on the powersuply (replace the IEC socket)

Peter

[Deleted]

Aug 30, 2009 9:28:14 GMT loupy31 said:

Hi All,
Just another quick question. As far as AC supply goes, Do I need to use a 3 wire lead from the wall socket to the unit (IEC), which would include earthing to the case/enclosure, Or can it be a 2 wire Ac cable into a 2 pin AC socket on the powersuply (replace the IEC socket)

Peter

Hi Peter
It would be safer to use the 3 wire lead . You really need fully insulated cases etc. if you don't use an earth wire. This would mean insulated on/off switch etc.
Alex

[robertkd]

Alrighty then, it's been a while so time for some more refinements from the house of Chimp Audio ;D

Currently settled on 6R8 emitter resistors and quiescent current of 35mA

Finished tidying up the horrid mess the silicon chip folks did to the op-amp input PCB layout as well as doing a decent job on ground plane around the Op-Amp and the decoupling has paid off in terms of transparency and fidelity.


AV=4 this gives a nice range for the switched attenuator and keeps it in the 9 o'clock to 3 o'clock sweet spot for the attenuator for my humble collection of headphones and musik consumption

I have opened the HF response using 33pf polystyrene's now arrives at an AV = 2 at ~ 3Mhz and gently slopes to unity way past radio Moscow on the short wave band May experiment on adding a few pf across the inv and noninv inputs just to be sure, to be sure, we've lost interest at maybe 20Mhz although it's not an issue so maybe if it ain't broke,... don't feck with it,...

Current limiting is totally transparent , it has a huge margin to cater for anything I have been able to throw at it including 22R load resistors, so it now doesn't care if you unplug the headphones with signal or not,.... happy vegemite

next I toying with taking the collector supplies to maybe +/- 22V

At the moment this is the closest thing to glass I have heard it's just lucid, and with the DAC there isn't 'black' between items it's 'void' between items

so here it is the continuing saga in the pursuit of audio Zen-ness

[Deleted]

Robert.
It's a shame there is approx. 1200km between you and Alex. Would love to know if you both prefered one version (ie agreed purely on sound produced).

Maybe you guys could post back and forth once yours is finalised the way you want it/as far you want to go. ;D

Chris

[Spirit]

I'd be interested to see Robert's take on a class A ;D



Radio Moscow indeed

[Deleted]

Oct 31, 2009 2:20:01 GMT Spirit said:

I'd be interested to see Robert's take on a class A ;D

Radio Moscow indeed

Phil
What we have seen with Robert's recent schematic is a huge decrease in the value of the FB capacitor from 1,200 picofarads to
33 picofarad. There is now a much greater degree of commonality
in thought,between our 2 designs.
Note that Robert has also vastly improved the PSU area by the use of the additional voltage regulators. This would undoubtedly bring it very much closer to the version using a JLH. The main advantage of the JLH, then becomes the ability to put it in line, without major surgery to the EXISTING PCB. This makes it much easier for the home constructor, without a lot of experience in this area.
I would be interested to hear Robert's subjective findings after such a dramatic widening of the HA's bandwidth. Not just how it sounds now, but how it compares with his original implementation.I suspect that both our versions would now be very much closer in perceived SQ.
I must admit that I would be interested to see Robert's take on a Class A HA, especially, if he used a LTP in the front end, and was game to try my front end balancing in combination with his own otherwise unique design. If he was game to do that, I would be happy to give him my remaining dual transistors. My balancing method was put to the test in late 2008 in DIYAudio's "Current Mirror" thread.
Alex

www.diyaudio.com/forums/solid-state/133018-current-mirror-discussion-15.html

[Spirit]

Oct 31, 2009 3:43:40 GMT @sandyk said:

I must admit that I would be interested to see Robert's take on a Class A HA, especially, if he used a LTP in the front end, and was game to try my front end balancing in combination with his own otherwise unique design. If he was game to do that, I would be happy to give him my remaining dual transistors.

So how 'bout it Robert?
I believe Will is (was? dunno what the B32S is driving) running his 701s with his Class A and loving it.

[robertkd]

Oct 31, 2009 2:20:01 GMT Spirit said:

I'd be interested to see Robert's take on a class A ;D

mmm Oh you mean class A as in like




Radio Moscow indeed

well I would just hate to have some AM short wave radio version of the nutcracker suite play over my new Missy Higgins album

[robertkd]

No seriously I haven't considered the Class A amplifier and for a number of reasons firstly the SC HA I have is sounding pretty good, well good enough for the time being. Also I'm not sure of the benefits to be had over SC HA in terms of preformance I mean seriously what cross over distortion if it's there, for me it's not an issue.

Would Class A be any better for linearity, THD and IM? Not necessarily and this isn't a dig at anyone it's just the reality of physics and solid state devices, of course the Class A topology may well have a different sonic signature, is it better or just different??

Here are a couple of other reasons I have one of these

www.motu.com/products/motuaudio/traveler-mk3

four of these and I'd like to upgrade them,..

www.akg.com/site/products/powerslave,id,1055,nodeid,2,_language,EN,pid,1055.html

A few other bits and pieces and of course a MBP to record on.

So that's way I haven't bothered with it yet,...

[Deleted]

Also I'm not sure of the benefits to be had over SC HA in terms of preformance I mean seriously what cross over distortion if it's there, for me it's not an issue. Would Class A be any better for linearity, THD and IM? Not necessarily and this isn't a dig at anyone it's just the reality of physics and solid state devices, of course the Class A topology may well have a different sonic signature, is it better or just different??

Robert
If you carefully read the reports from people who have constructed the Class A HA , preamp or Power Amp, you will find that the major noted improvements over the SC HA come when the very closely matched LTP transistors have virtually identical collector current through them. i.e. close to 0mV difference between their collectors. This ensures that they thermally track each other , and the VBEs which are already close together, remain so. Hugh Dean (AKSA from DIYAudio and Aspen Audio,Australia) said this :
The transfer function of the LTP is appalling, resembling a slewed 'S'. Only the narrowest portion is linear, and I believe it is the non-linearity at this input stage to the amp which is responsible for robbing much of the detail required to deliver good imaging and resolution. Consider this.

An amp has an open loop gain of 95dB, and a closed loop gain of 30dB, giving a feedback ratio of 65dB, which is 1780. If the output of the amp is 40Vpp (25W into 8R), then the differential voltage appearing across the two LTP bases will be 40/1780, which is 22.5mV. I don't need to tell you that with Vbe figures of 620mV or so, and maybe a deviation between two well matched devices of no more than 5mV, this 22.5mV differential severely unbalances the LTP, taking it outside the very narrow linear region of the S curve, and causing quite high distortion. What is worse is that as the LTP is unbalanced, one device pass more current, increasing it's Vbe, while the other pass less, decreasing its Vbe, and this is positive feedback, causing even more distortion. Anything which can bring the operative devices of a LTP into constant Vbe would improve things, and your balancing of both current and voltage is probably doing this ..."

[robertkd]

Oct 31, 2009 9:11:10 GMT @sandyk said:

Also I'm not sure of the benefits to be had over SC HA in terms of preformance I mean seriously what cross over distortion if it's there, for me it's not an issue. Would Class A be any better for linearity, THD and IM? Not necessarily and this isn't a dig at anyone it's just the reality of physics and solid state devices, of course the Class A topology may well have a different sonic signature, is it better or just different??

Robert
If you carefully read the reports from people who have constructed the Class A HA , preamp or Power Amp, you will find that the major noted improvements over the SC HA come when the very closely matched LTP transistors have virtually identical collector current through them. i.e. close to 0mV difference between their collectors. This ensures that they thermally track each other , and the VBEs which are already close together, remain so. Hugh Dean (AKSA from DIYAudio and Aspen Audio,Australia) said this :
The transfer function of the LTP is appalling, resembling a slewed 'S'. Only the narrowest portion is linear, and I believe it is the non-linearity at this input stage to the amp which is responsible for robbing much of the detail required to deliver good imaging and resolution. Consider this.

An amp has an open loop gain of 95dB, and a closed loop gain of 30dB, giving a feedback ratio of 65dB, which is 1780. If the output of the amp is 40Vpp (25W into 8R), then the differential voltage appearing across the two LTP bases will be 40/1780, which is 22.5mV. I don't need to tell you that with Vbe figures of 620mV or so, and maybe a deviation between two well matched devices of no more than 5mV, this 22.5mV differential severely unbalances the LTP, taking it outside the very narrow linear region of the S curve, and causing quite high distortion. What is worse is that as the LTP is unbalanced, one device pass more current, increasing it's Vbe, while the other pass less, decreasing its Vbe, and this is positive feedback, causing even more distortion. Anything which can bring the operative devices of a LTP into constant Vbe would improve things, and your balancing of both current and voltage is probably doing this ..."

What did I miss something, I was politely saying I don't give a feck about it at this time.

[Deleted]

I was politely saying I don't give a feck about it at this time.

Robert

Perhaps if you had not included this sentence,""it's just the reality of physics and solid state devices" I would have clearly understood your complete lack of interest, and not replied further ? As I will now do.
Alex

P.S.
How are your knees coming along after the operations. Are you slowly becoming pain free at last?

[robertkd]

repudation what? please,.. I said nothing derogatory at all, mealy indicated we are getting to the levels that our current technology can yield. Whether we find that in class a tube, solid state, class D or class ABx it doesn't matter.

I will say it again "it's just the reality of physics and solid state devices"

I'm surprised you didn't pick the oxymoron "solid state"

[Will]

Robert, if I remember correctly, you have a pair of AKG K271mkII (headphone stash thread). How do you find them, and how do they compare to your K702?

I ask as I'm after a pair of circumaural closed back headphones, and as I like the sound of my K701, and the K271 are easy to come by on eBay.
I'd be interested in your thoughts.

Ta

[robertkd]

Oct 31, 2009 16:08:53 GMT Will said:

Robert, if I remember correctly, you have a pair of AKG K271mkII (headphone stash thread). How do you find them, and how do they compare to your K702?

I ask as I'm after a pair of circumaural closed back headphones, and as I like the sound of my K701, and the K271 are easy to come by on eBay.
I'd be interested in your thoughts.

Ta

mmm sticks tip toes in the water, thinks someone asking my opinion,...

Well now that you ask, mine are the k271 not the current k271 mk 2. They are not k701's

They are a nice set of cans but they are not my favourite closed back circumaural I like the way the k701's sit around my ears, I don't like the way the k271's sit.

Honestly it depends what your looking for Mike likes the HD681, but I've not heard them personally,...

I prefer the style of k701 and that of Audio Technica AD1000's style, so for closed back types and not in preference order, damn it their all my children ;D

AKG k271 Nice fairly natural, more typical AKG sound
AT A500 Nice but think maybe go the A900's
AT ATH A55 Great bang for the $ (eh Pound)
AT M50 solid performer maybe a bit bass heavy

Again there not chalk and cheese differences as you can appreciate cans can be very personal, of the above the k271, A500, A55 are more or less similar with the M50 having a definite sonic signature. The A55 and M50 perhaps offering the best isolation for external noise.

Say hasn't Mike got a pair of W1000's for sale,...

[Deleted]

Say hasn't Mike got a pair of W1000's for sale,...

Will
That is a great suggestion from Robert. The W1000 don't have the butt kicking low end that Mike likes, but they are excellent headphones, and not just because I have a pair
I have heard TonyG's 701s,in comparison with Tony's own W1000, but we both still felt that the W1000 had the edge with a more natural sounding decay with drums, as well as in a few other small areas. Not too much between them otherwise though, other than perhaps personal preference.
Alex