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 154155156157158 345>
Author
  Topic Search Topic Search  Topic Options Topic Options
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: 22 Sep 2021 at 3:39pm
The great difficulty in adding a voltage regulator is the voltage loss. At full power, the ripple voltage on the smoothers is 4.8V, and the TL783 dropout voltage should be approx. 4.5V for the small current required. Additionally, I have to waste a few mA's because of its minimum current (15mA).

By using a tapped transformer primary, set to the correct mains voltage Wink, the full load voltage at the smoothers should be 71V5 (52V secondary).

The maximum obtainable voltage will be 62V2. I shall aim for 62V.

The VAS emitter voltage is comparatively high, because of the two transistor configuration I decided to stay with. I estimate 8V worst case.

That leaves us with 54V, and then we must minus the OP stage base spreader at 2V5, so we only have 51V5 to centre the signal swing about.

That gives 25V75 peak, which is 18V025 rms. It should just scrape 40.6 watts.

I think I could get away with letting a little ripple through, as it would only occur at >40.6 watts. So, I might just set it for 65V. Or, I could set it half way between 62V and 65V to cover component tolerances.

Calculator time!
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: 22 Sep 2021 at 7:23pm
Don't you just love equations like

Vo = Vref  ((R2+R1)/R1)

All you want to know is the value of R2 to give the voltage you want, but they never show you the equation for that.

So, knowing the value for R1 from the datasheet, and that Vref is 1.25, let's rearrange it.

I want 63.5 volts, so by plugging in numbers, we can end up with R2, which I've called X.

Vo = Vref  ((R2+R1)/R1)

63.5 = 1.25 ((x + 82)/82)

63.5/1.25 = (x + 82)/82

63.5/1.25 x 82 = x + 82

(63.5/1.25 x 82) - 82 = x

= 4083.6 ohms

Now, to prove the rearrangement...

Vo = 1.25 ((4083.6 + 82)/82) = 63.5
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: 22 Sep 2021 at 8:27pm
The next problem is that a 4083.6 ohm resistor - if there was one - would get hot with 63.5V - 1.25V across it.

With 0.6W rated metal film types, the maximum wattage ought to be 0.12W for it to run cool.

so, 0.12/62.25 = 1.93 mA, and

62.25 / 1.93 = 32k25, nearest is 33k, but you'd need 8 in parallel, and that causes a PCB tracking problem.

But you can put 8 in series, and daisy chain them where they'll fit, and the nearest value is 510 ohms, giving 4080 ohms.

So, the final voltage is 63.45V.

The next problem is that the bootstrap wants to lift the power supply by the peak output voltage minus the voltage dropped by R14 || R20, and if we reckon the VAS current is 7mA, they drop 8.4 volts.

By accident, the amp might do 50 watts (into 8 ohms), which is 28.28V peak, so -20V is dropped across 1.2k = -17mA rounded up.

So, if we want 2.5mA for the IPS, and the minimum TL783 current is 15mA, totalling 17.5mA, we must add 17mA, making 34.5mA. That's roughly the forward current I had used in transient simulation to get the slew rate to where it's supposed to be.

Now, if the TL783 drops 9.3 volts to cover ripple and its own dropout voltage, the watts it must dissipate are 9.3V x 0.0345 A = 0.32W.

The junction-to-ambient thermal resistance is 25.3 C/W, so the increase will be 25.3 x 0.32 = 8C, above ambient. And if ambient inside the case is 40C, the junction temperature will be 48C. The TL783 will feel hot, but won't melt solder.

We can risk 60C and give it some copper to sink the heat, so we can go to 20C above ambient, and 20 / 25.3 = 0.79W. The maximum current would therefore be 85mA.

So, we need to be between 35 and 85 mA.

Confused? I'm trying not to be.
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 (1) Thanks(1)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 22 Sep 2021 at 9:35pm
Let's try using power resistors to save space and go for the more serious current I think we might need to overcome the bootstrap.

We want 15mA minimum for the device to regulate, and we want 17mA to counter the bootstrap, so the device reference resistor might as well consume that, so the minimum current across it is 32mA.

Now, the bootstrap will also pull down 28.28 volts peak over 1.2k, which is 23.6mA, and it must come from the VAS rail (the regulator). With no signal, it is just the 7mA quiescent.

And then there is the 2.5mA for the IPS.

So, 32 + 23.6 + 2.5 = 58.1mA (58mA for ease of manipulation).

The TL783 voltage drop is 9.3V and so 9.3 x 0.058 = 0.54 watts. The temperature rise is 25.3 x 0.54 = 13.7C, and if we aim for a 60C limit, the internal ambient must be a maximum of 46.3C. That's feasible as it will contribute heat to the inside of the case, as well as the output stage and transformer.

So, the reference voltage is 1.25V and it needs 32mA to flow, so 1.25 / 32 = 39 ohms - a "prefered" value.

From (Vo/Vref x R1) - R1 = R2, (63.5/1.25 x 39) -39 = 1942.2 ohms.

Checking back, Vo = Vref  ((R2+R1)/R1), 1.25 ((1942.2 + 39)/39) = 63.5V

Can we use two 1k resistors?

1.25 ((2000 +39)/39) = 65.35V, perhaps a bit too much ripple breaking through? OK, worth a try.

So I need two 1k 5W resistors...

Damn!
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: 23 Sep 2021 at 9:25am
Now, by adjusting values we can use 36 ohm for R1 and 1k8 for R2, and make R2 small and cool by using a TO-220 power film resistor, rated at 35W. And I will have to wait until tomorrow for its delivery. Unhappy
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 (1) Thanks(1)   Quote Graham Slee Quote  Post ReplyReply Direct Link To This Post Posted: 24 Sep 2021 at 4:51am
Half of this topic's last three years has been spent avoiding each other in the HiFi System Components team - either by enforced lockdown or just plain common sense. In fact, we'd already distanced into our solo careers just in case the 2015 hand of fate tried its ruinous trick again. I had become so used to having my two pals working with me during the development of the Proprius, the Bitzie, the Majestic, and the MM Accession, because there was always somebody around to bounce ideas off, and to audition each stage of development. But since 2015, and even more so since the lockdowns, it's been me working in isolation.

I'd become lone designer and office lad. They'd become flexible sub-contractors just in case the suspicious leeches at HMRC are wondering. I'd managed the Accession C and PSU1 Enigma, but at least I could send John and Leo samples to play with, and to work from.

The amplifier I've been working on here has been a lot more like the DIY of my earlier career, working on it when I can, other things allowing, but with the trappings of a real company - an office, workshop and most importantly, an Audio Precision AP525 analyser - not forgetting my SPICE simulator.

Even so, there have been times where some concepts have been hard to grasp. I'm sure you're all fed up of me telling you how little was documented about vintage transistor amps, and that I've been trying to put flesh onto these old dinosaurs, to help you and me understand them.

Then we got Skype, so John and I could virtually and occasionally work together, often figuring out customer enquiries, making sure we bought enough components during shortages, having the odd laugh - that sort of thing. We sometimes got a virtual visit from our US loaner program friend, Bruce Kohl (BAK) from Kentucky. I don't know about Zoom, but Skype has been great.

Bruce and I had a simultaneous interest in these old amplifier designs, and still do. He's rebuilding an old Dynaco ST120, and we'd previously communicated by email, but Skype made a huge difference. The great thing about Skype is that even when the bandwidth goes down, we can still type - even if it's "hunt and punch" as Bruce calls it.

It has been really good to be able to bounce ideas and concepts off somebody else, especially as I discovered Bruce's background has been medical and military electronics. Often we've typed into the small hours, and been able to adapt formulae and equations meant for today's amplifier circuits, and make them fit our retro projects. It has greatly helped our understanding of many a vintage transistor amp, and I think we must have discovered a number of pitfalls in some products (our lips are sealed), and we've been able to get into the minds of quite a few old school designers (hope they don't mind).

So this is why I'm writing this post, to show my gratitude to my fellow old timer, whose spent a lot of time with me on this journey. He's corrected me - I've corrected him - sometimes we disagreed only to realise we were talking about the same thing. It's a marvellous thing being able to get so far just by "hunt and punch" (on the keyboard!). As for the face to face, yes, it's good if you like highly pixelated images and muffled sound, but it's a long way between Yorkshire and Kentucky!

Today I hope to be able to make the voltage amp regulator, and would like to thank Bruce for the pages of calculations, the checks and balances we must have done, in just the last two days, to try and get it right.
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: 24 Sep 2021 at 3:12pm
Graham wrote: "Today I hope to be able to make the voltage amp regulator, and would like to thank Bruce for the pages of calculations, the checks and balances we must have done, in just the last two days, to try and get it right."

 You still have to build it and see! Wink

 Remember there are variables in all parts, even 1% tolerances ones... and the sum of all the part variables cannot be calculated. The prototype building and empirical testing are the only way to truly verify a design.

 As engineers like to calculate everything to get close to a design that works before building it,
from bad experiences, they learn "it can still go POOF in a cloud of smoke"!Embarrassed

 I showed to Graham some of the many hats I've worn... real hats.

 In the late 1960's, I "cut my teeth" in engineering by designing tube circuits for audio, video, and radio. 
The medical and military stuff came later... after working on all consumer electronics.
 (Some turntables I've repaired were the worst designs mechanically.)


Edited by BAK - 24 Sep 2021 at 3:37pm
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
 Post Reply Post Reply Page  <1 154155156157158 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 0.203 seconds.