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 3233343536 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: 03 Dec 2018 at 3:31am
Oh, and now there are chips which means the transistor manufacturers don't anymore. And with attitudes like "I don't care so long as it sounds good", we eventually reach here...

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: 06 Dec 2018 at 6:07am
Fools Errand

The designer can rely on op-amp simulation models telling the truth (?), but transistors? Well, manufacturers probably think it's enough to recommend particular types for particular jobs, and the DC conditions and transition frequency plus capacitance are all quoted on the data sheets, so what the heck?

But simulation saves pages of calculus (for those who can do it...), or more rudimentary calculations for the likes of me, but there is always nagging doubt about forgetting, misplacing, or sometimes miscalculating. So transistor models would help, but often relying on them is a fools errand.

And here I have an axe to grind with Motorola and whoever else mucked about (they will know who they are). The models of the MPSA06 are fool's errands! Designers know about a thing called "the sanity check" but work pressures often mean there isn't the time. And trying to measure what's going on at upwards of 10Mhz in a power amplifier run on unregulated DC...

It turns out that the bright sparks who have modelled the MPSA06 think it will continue beyond the frequency of light! I know it's good but come on! It is a good job that I'm using them in a design which can be checked by other means... my ears!

So what I said about my last tweak on stability should be ignored. The dominant pole shall revert to where it was! And with the addition of the smaller capacitor to deal with T1's gain contribution. The more "global" arrangement falls flat on its face, but at least it didn't burn-out the amplifier. Oh, and remember how well THD measured? But there's another mechanism at work here that I don't have the time to investigate.

I have always been suspicious of the designs of John Linsley Hood (deceased 2004). Many rated him highly and still do, but in my opinion, some of his circuits employed sticking plaster techniques, and his 1978 Hi-Fi News amplifier design featured the non-dominant pole compensation I refer to above. The explanation given is it removes slew-rate limitations. Well, the circuit does its "damage" which is "put-right" by the global negative feedback...? How many times have I read that global negative feedback is not the panacea? Perhaps the names JLH and Marcus Scroggie ("Cathode Ray") should not appear in the same sentence...?

So, modelling with a number of other transistors I found the source of this undesirable sound quality - the instability - and it should have appeared obvious, but I did it, and now I have to go and put it back!
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: 06 Dec 2018 at 8:05am
Artists impression of case for power amp version...



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

Joined: 14 Mar 2010
Location: Kentucky, USA
Status: Offline
Points: 1744
Post Options Post Options   Thanks (0) Thanks(0)   Quote BAK Quote  Post ReplyReply Direct Link To This Post Posted: 06 Dec 2018 at 10:39am
Originally posted by Graham Slee Graham Slee wrote:

Fools Errand

The designer can rely on op-amp simulation models telling the truth (?), but transistors? Well, manufacturers probably think it's enough to recommend particular types for particular jobs, and the DC conditions and transition frequency plus capacitance are all quoted on the data sheets, so what the heck?

But simulation saves pages of calculus (for those who can do it...), or more rudimentary calculations for the likes of me, but there is always nagging doubt about forgetting, misplacing, or sometimes miscalculating. So transistor models would help, but often relying on them is a fools errand.

And here I have an axe to grind with Motorola and whoever else mucked about (they will know who they are). The models of the MPSA06 are fool's errands! Designers know about a thing called "the sanity check" but work pressures often mean there isn't the time. And trying to measure what's going on at upwards of 10Mhz in a power amplifier run on unregulated DC...

It turns out that the bright sparks who have modelled the MPSA06 think it will continue beyond the frequency of light! I know it's good but come on! It is a good job that I'm using them in a design which can be checked by other means... my ears!

So what I said about my last tweak on stability should be ignored. The dominant pole shall revert to where it was! And with the addition of the smaller capacitor to deal with T1's gain contribution. The more "global" arrangement falls flat on its face, but at least it didn't burn-out the amplifier. Oh, and remember how well THD measured? But there's another mechanism at work here that I don't have the time to investigate.

I have always been suspicious of the designs of John Linsley Hood (deceased 2004). Many rated him highly and still do, but in my opinion, some of his circuits employed sticking plaster techniques, and his 1978 Hi-Fi News amplifier design featured the non-dominant pole compensation I refer to above. The explanation given is it removes slew-rate limitations. Well, the circuit does its "damage" which is "put-right" by the global negative feedback...? How many times have I read that global negative feedback is not the panacea? Perhaps the names JLH and Marcus Scroggie ("Cathode Ray") should not appear in the same sentence...?

So, modelling with a number of other transistors I found the source of this undesirable sound quality - the instability - and it should have appeared obvious, but I did it, and now I have to go and put it back!
 You had to "turn that stone over" or it would have bothered you.
Again, showing you always explore all possibilities to make your design sound its' best.
 
So, the position of the dominant pole capacitor C7, 47pF, is put back to where it was on page 23. And C5, 10pF, is also put back.

Is this correct?....
 C7 sets the gain-bandwidth product 0dB-crossing frequency (GBP) and 
C5 reinforces  the stability above the GBP while having very little affect on the GBP.
 And, C5 gives additional control of the gain at frequencies above GBP
(by making T1 "behave")?
 


Edited by BAK - 06 Dec 2018 at 11:06am
Bruce
AT-14SA, Pickering XV-15, Hana EL, Technics SL-1600MK2, Lautus, Majestic DAC, Technics SH-8055 spectrum analyzer, Eminence Beta8A custom cabs; Proprius & Reflex M or C, Enjoy Life your way!
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: 06 Dec 2018 at 8:43pm
Good question.

You have T2 gain stacked on top of T1 gain so C7 acts on T2 leaving T1 alone you would think, but in reality C7 is local NFB making T2 base virtual earth at high frequency so shunting T1 gain, so it does have some effect on T1 after all.

But it doesn't continue much lower than that, and T1 gain causes overall gain to flatten out before gain margin reaches a safe point (circa -10dB), and also phase margin is a little shallow.

C5 in reality will only take the overall picture to unity, but remember C7 is acting on the product of both T1 and T2 because of the virtual earth coupling, and combined, take gain margin below -10dB into the stable region.

But C5 (as you mention) has to act higher in frequency so as not to "collide" with the dominant pole curve, which can cause abrupt phase reversal (although that is below unity gain), so is set at a smaller value.

The result is an orderly 6dB/octave roll off from C7 to just below unity crossover, where C5 joins in and steepens the curve to 12dB/octave, rapidly approaching -180°, but diving below -10dB before it gets there.

In fact the gain dips to near -20dB before recovering to around -10dB which according to the model finally decays around 5GHz, but parasitics should finish it off much lower.

First image shows margins with 47pF and 10pF. PM is 80 approx; GM is -14dB rising to just under -10dB.


Second image shows margins with 47pF only. PM is only 62 and GM is about -8dB.


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

Joined: 14 Mar 2010
Location: Kentucky, USA
Status: Offline
Points: 1744
Post Options Post Options   Thanks (0) Thanks(0)   Quote BAK Quote  Post ReplyReply Direct Link To This Post Posted: 07 Dec 2018 at 1:44am
Is the MPSA06 satisfying your design for the IPS and the VAS?
Bruce
AT-14SA, Pickering XV-15, Hana EL, Technics SL-1600MK2, Lautus, Majestic DAC, Technics SH-8055 spectrum analyzer, Eminence Beta8A custom cabs; Proprius & Reflex M or C, Enjoy Life your way!
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: 07 Dec 2018 at 8:27am
Originally posted by BAK BAK wrote:

Is the MPSA06 satisfying your design for the IPS and the VAS?


Shall we say thus far they're OK?

D Self measured many suitable VAS transistors and found it produced the least distortion. I don't have the test lab he has so I'll just have to believe him.

I found a SPICE model for it over at Central Semiconductor which seems to be sane, and used it for the charts above. I also modelled using BC337's which gave similar results, but these won't work in the real amplifier as their sustaining voltage is too low. I tried the 2N5551 model but it reports a higher open-loop gain so I guess it's faulty.

I'm running it in whilst I work on the printed circuit artworks, not playing music but listening to the voices from the UK Parliament on the latest thrilling instalment of the demise of the United Kingdom. I find listening to all those people getting it wrong helps me concentrate on getting my layouts right LOL

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 3233343536 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 0.141 seconds.