New Posts New Posts RSS Feed - 1970s Design Indulgence
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Welcome to the Graham Slee Audio Products Owners Forum

 

Open to all owners plus those contemplating the purchase of a Graham Slee HiFi System Components audio product and wishing to use this forum's loaner program: join here (Rules on posting can be found here)

This website along with trade marks Graham Slee and HiFi System Components are owned by Cadman Enterprises Ltd


1970s Design Indulgence

 Post Reply Post Reply Page  <1 5960616263 345>
Author
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 14 May 2019 at 8:04am
Thank you.

If I get the chance I'm phoning Trans-Tronic and asking them if they made this Amplimo transformer.

It doesn't buzz! It did at switch-on, and with the computer noise I couldn't hear it after that, but switching computers off last night I didn't hear the buzz which every single RS transformer did. Transformer buzz is due to large magnetising current (not DC on mains) usually due to inferior core material. This transformer cost me £8 more than the RS equivalent. I'd rather pay a little more to get something which works properly RS!
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
peterb View Drop Down
Senior Member
Senior Member
Avatar

Joined: 02 Feb 2017
Location: Cornwall
Status: Offline
Points: 332
Post Options Post Options   Thanks (0) Thanks(0)   Quote peterb Quote  Post ReplyReply Direct Link To This Post Posted: 14 May 2019 at 9:32am
Am I allowed to say, 'told you so' !!!
Trans-Tronic make good products and have a high technical ability, both in design and manufacturing. I had a very uneventful experience of their products when I was working with them.
Peter
---------------------------------------
Dual 505-1, Cyrus CD T, DIY 80W MosFet amp and PreAmp, 2xKEF 103.2
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 14 May 2019 at 12:56pm
I'm quite happy to be told so Smile

However, as a design indulgence I think it is valuable to understand how these things affect the sound, and what mechanisms drive that. There is very little written to help the layman, never mind the experienced designer, when faced with such problems.

I have been on sites which ridicule the existence of these artifacts, and I'm talking about engineers I have respect for. Maybe they consider long-term always-on testing silly too?

I don't think designers and DIYers should be left in the dark, and for facts of physics to be hidden. OK, these "facts" might not even be known, but they should be searched out and made available.

FWIW, I listened (read actually) the views of electrical engineers on the workings of toroidal transformers, including that when idle or drawing little current, the magnetising current and hence core saturation increases. The waveform of which isn't very sine. These distortions creep in to the power supply and hence output signal, but as they do not conform to recognised measurements, they're never seen on the analyser.

Capacitors change character in ways we have yet to learn, and it seems even a little heat on the leads of what one would think is a "burned-in" component, changes its character again, only to resume its usual character just when you think you got it right!

Apply that heat to your finger and you'll soon scream ouch! When we apply it to a manmade plastics film do you think it results in no apparent difference to the structure? Just because a £10,000 analyser cannot "see" it doesn't mean it didn't happen, it's just because (as JLH said) we haven't bothered to investigate further.

I have talked with two industry measurement "experts" and they agree we haven't really scraped the surface, but you have to keep quiet about it so you're not lauhged at or worse.

That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
BackinBlack View Drop Down
Senior Member
Senior Member
Avatar

Joined: 05 Feb 2012
Location: Hinton, N'hants
Status: Offline
Points: 2020
Post Options Post Options   Thanks (0) Thanks(0)   Quote BackinBlack Quote  Post ReplyReply Direct Link To This Post Posted: 14 May 2019 at 2:11pm
No laughing here, just a knowing smile.
It's really gratifying to read and learn from your experience and research into the way these modern components behave as they age.
As for burn in, I remember from aeons ago (early 60s) replacing failed condensers and valves in my great uncles wireless. Freshly recharged accumulator and new HT/Grid batteries and all was well again, but after thanking me he did comment that it sounded very "bright", a twiddle of the tone control sorted that. A few weeks later he reported that it all sounded just like before and the tone control was back to normal.
A case of burn in perhaps?

Ian
Just listen, if it sounds good to you, enjoy it.
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (1) Thanks(1)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 15 May 2019 at 8:29am
Oh, and the other reason for trying a stock off-the-shelf transformer was that I, being very much like the amateur constructor, don't have many other options available.

You might be surprised by this, after-all being a manufacturer you'd think I'd have suppliers only too willing to supply me.

I first approached Trans-Tronic's Philip Pickering circa 1988 some 10 years into his "journey" when he was on Sheepsbridge industrial estate. He designed me a special output transformer for use in a fire-evacuation amplifier, to drive an impedance of 0.6 ohms.

The project for Cable & Wireless subsidiary Davis, was to design and supply 120 of these for installation in all their contracts such that in an emergency the fire chief could talk to every resident and get them out of the building safely.

Therefore, my first dealings with Trans-Tronic, 31 years ago, was quite substantial for Philip.

Since then I introduced others where I worked to Trans-Tronic, which resulted in further substantial orders for them.

And since starting the present business, I have bought somewhere in the region of 20,000 transformers from them. As well as inductors and potted items.

But Philip retired, and today Trans-Tronic are too busy for the likes of me.

I have a similar difficulty with a metalwork manufacturer who used to supply me with £60,000 of custom metalwork every year, representing 1/10th of their turnover. As a result I can only spend £40,000 a year with them...

Is there any wonder why British manufacturing is in such decline?

That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 19 May 2019 at 7:39am
It would seem that the amplifier itself is not a problem, especially as, being based on the Proprius, itself being based on H C Lin's design; the Proprius suffers none of these ills.

Therefore the power supply is the problem. So do we resort to making an amplifier with additional gain to be used up as negative feedback as a sticking plaster?

And is this the reason why solid-state is such a problem, and that valves are that much better?

Or is it that valve designs consider the power supply in much greater depth?

I think you will find it is the latter. I would advise solid-state designers to study valve amp design. I know I have, and I still do.

The Proprius was intended all along to see-off compact class-D designs and its switched mode power supply was an early design decision. It led to the Proprius keeping its great sound full time when left on.

We have a choice when using transformers to use the old fashioned EI transformer, or a modern toroidal. To keep hum down, this design went the toroidal direction. It might have been a mistake.

A toroidal transformer is far from being perfect. How an amplifier manufacturer can simply stick in a 230V 50/60Hz toroidal transformer and expect it to deliver the same results on voltages up to 250V on 50Hz, is beyond belief.

The reason for me designing audio electronics in the first place was not as a job or career choice, but because I wanted to hear better than what was available commercially. Then having done so, it seemed obvious that I should follow that path as a career.

OK, down to the technology bit...

A 230V 50/60Hz toroidal transformer will work best at 230V 60Hz (because if it didn't it would not do as advertised). At 50Hz it will saturate more than at 60Hz. And that begs the question why make it for 60Hz when on 230V the frequency is 50Hz?

It will also saturate much more at 250V, and 250V can be expected even when it's supposed to be 230V. The wording changed but the generators never did!

Have you ever heard the comment that solid-state sound best at high volume? Ever wondered why? At high volume more current is drawn from the transformer and so its saturation will drop.

So, it is transformer saturation which is responsible for the solid-state sound, plus the sheer amount of negative feedback used as sticking plaster. The Proprius doesn't sound "solid-state".

So, can it be fixed? My opinion is that a toroidal transformer should be made area specific. If the voltage is 250V and frequency is 50Hz, then it should be made for that. It does mean problems for export, but why try to sell something into areas with different mains supplies when you know the product cannot perform? It brings reviews into disrepute doesn't it?

The Amplimo transformer (made by Trans-Tronic) is better engineered but still "grumbles" on 250V. It is after all, a 230V transformer.

When saturation happens the waveform isn't a nice 50Hz, it is distorted, and therefore contains other frequencies. The rectifier and reservoir capacitors designed for 50Hz (or even 60Hz) actually have to deal with harmonics, but can they?

There is also the switching of the rectifier which transitioning to its off time will generate inductive spikes in the transformer secondary, which nobody seems to talk about (except Morgan Jones), but must be obvious?

Snubbing the transformer secondary has led to a sonic improvement, but obviously any gains here are diminishing because the hard worked capacitor eventually settles into its routine and is no longer "new". Values used: 220n "X-cap" in series with 1k.

Spikes will reappear if not to the same extent as not snubbing, so where do we look next? These spikes if they were "a frequency" are not compatible with a large electrolytic reservoir capacitor because of its flat bottom impedance curve as it goes from purely capacitive to inductive.

Here we can bypass using an "X-cap" (mains repetition rated) across the electrolytic's terminals, and trust the spikes are reduced. Value used: 220n "X-cap".

What escapes might be helped by also connecting a smaller electrolytic across the terminals, and I used value: 100uF/100V.

Also note I am not using bridge rectifier diode snubbers as the simulation showed they made matters worse.

Doug Self explains that a capacitor subjected to repetitive and continuous "abuse" such as this needs to be of a type intended for it, and so to use a capacitor intended for continuous across-mains use.

Morgan Jones (Valve Amplifiers) illustrates where the spikes are going to happen, and this is substantiated by my own simulation modelling, and as Morgan Jones explains, the impedance must be low straight after the rectifier, suggesting a similar regime.

And in fact, Morgan Jones extends the argument to include a "swinging choke" (inductor) between rectifier and reservoirs (smoothing), such is the care necessary with valve amplifiers.

For now I am not really wanting a choke the size of the mains transformer itself, so I am making use of a second wave of filtering in the form of the 0R22 power resistor between first and second reservoir capacitors - this being a Bob Cordell idea.

I am hoping by now that the technical reader of this marathon work will be realising that the one cap per rail commercial solid-state amplifier cannot be as good as the reviews make out - it should be obvious!


Edited by Graham Slee - 19 May 2019 at 8:11am
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
Graham Slee View Drop Down
Admin Group
Admin Group
Avatar
Retired

Joined: 11 Jan 2008
Location: South Yorkshire
Status: Offline
Points: 16298
Post Options Post Options   Thanks (0) Thanks(0)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 20 May 2019 at 10:17am
What causes brightness?

If an amplifier measures flat - which it does - then all frequencies should be in balance as determined in the recording.

So, if a frequency or bunch of frequencies appears to be louder than others, it can only be that their harmonics are louder than the harmonics of other frequencies.

The only way these harmonics can be louder is if distortion at those frequencies is high.

It is normally sibilance which is most annoying. The sibilance frequencies are normally 3kHz to 9kHz, so the second harmonics of these are 6kHz to 18kHz, and the third harmonics are 9kHz to 27kHz. It is usually third harmonics which are objectionable.

In the region 9kHz to 27kHz (which also takes in some of the second harmonics) the amplifier's open loop gain is falling, and so we don't have as much distortion reducing negative feedback to try and cancel it.

Perhaps if we increase the amplifier's open loop gain it will help? By shunting some of T1's emitter resistance it can be increased to 77dB, some 14dB (or 5 times) more than what we had.

The open loop frequency response goes to 36kHz, and so by Otala, we can see its high enough not to cause too much transient intermodulation distortion. But it will not be stable so we have to compensate it.

The compensation we had is insufficient because open loop gain has increased 5 fold, and T1 doesn't have as much degeneration. Whereas 47pF would do before, we must have 5 times that to maintain stability. So, 220pF should be the value which works.

We had made use of global negative feedback compensation but now T1's emitter padding resistance has been bypassed this doesn't work properly, so compensation has to go back to T2's base - in effect Cdom returns.

Cdom, because it is negative feedback, helps reduce high frequency distortion, so can we make it reduce some in the output stage? We can connect it from the output mid-point instead of from T2's collector, but only if we put some resistance in series with it, and 30 ohms simulated best.

So now we have 220pF in series with 30 ohms, output mid-point to T2's base, but we still need local compensation, as we always have in this design, between T2's collector and T1's base; 33pF still being OK.

This compensation by itself gives 34dB distortion reducing high frequency negative feedback at 20kHz. With gain set at 27dB and 77dB being open loop, the overall negative feedback or loop gain is 50dB, so we see that at 20kHz the purely resistive negative feedback is 16dB (16 + 34 dB being 50dB).

Either way, we have 50dB NFB to reduce second and third harmonic distortion of the sibilance frequencies, so the brightness should now be gone.

However, we have a rather large electrolytic capacitor as VAS (T2) emitter bypass which allows T2 to achieve its full gain, but will it do that cleanly between 9kHz and 27kHz?

Looking at impedance curves for electrolytic capacitors we see a bathtub curve which is due to the capacitor becoming more inductive as frequency increases.

We need T2's emitter impedance to be perfectly linear, but the bathtub curve shows this isn't the case. Bypassing the electrolytic is therefore essential. If we take 9kHz as being a possible problem for the bathtub curve (as we do not have exact data), and we are bypassing a total resistance of 1k ohms, the capacitor value works out at 18nF.

However, this is where it starts to take effect. If we use a value ten times greater it might reduce impedance sufficiently, indicating a bypass capacitor of around 220nF. In error I had used a value of 10nF, which was copied from the Proprius design, which uses a smaller electrolytic having a better bathtub curve.

To ensure it is completely polished-off, we can make the capacitor value even larger as long as it's resonant frequency is far and beyond the power amplifier's gain bandwidth product, which is around 2.5 to 3 MHz.
That none should be able to buy or sell without a smartphone and the knowledge in how to use apps
Back to Top
 Post Reply Post Reply Page  <1 5960616263 345>
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.01
Copyright ©2001-2018 Web Wiz Ltd.

This page was generated in 0.199 seconds.