Friday, August 10, 2007

Power supply voltage in a Quad 33; 12, 15 or 16V? Modification december 2010


The discussion started with the first article in the Netherlands about the revision/upgrade and modification of a Quad 33 and 303 by Ruud Janssen. See appendix 1

This upgrade, apart from the standard replacement of all the electrolytes, was based upon the increase of the power supply voltage to 16 volt, reducing the gain of the output amplifier stage, to cope with high level inputs, and increasing the gain of the PU amplifier to compensate for the lower gain of the output stage. Also the power supply voltage for the PU amplifier was kept at the same level by changing R300. Although the series resistors in the PSU are decreased when you install a 16V zener, this is still one of the disadvantages of this type of regulation, a PSU with a high internal resistance. This gives poor load regulation and higher crosstalk figures.

Considering the input signal levels, I built 33’s with 7812 and 7815 regulators, (check the remarks about the 7812/7815 below) There is more headroom with a 15V supply, correct. So with certain inputs, there is an advantage.
There is a big difference between the standard (12V) zener based supply and the regulator based supply, but the regulator based supply is difficult to sell in a DIY upgrade kit. Experienced DIY users can do this job, also it can be done in one of our shops.

The reason for keeping the PSU voltage at the same level for the PU amplifier was the following: To preserve the noise level produced by the first BC109 transistor on the PU board, Ruud Janssen stated that the DC bias current should not change. This looks like a valid argument, but increasing the gain of the PU amplifier to compensate for the lower gain of the output amplifier increases the noise level of the PU stage.

The decrease of the gain of the output amplifier is no guarantee that a high level output from a CD player or other source will not overstress the input amplifiers on the tape adapter board. All the input signals, except the tape input are routed via the input amplifier on the tape adapter board. This is even more a consideration when someone uses the modified PU input selector board for connecting a CD player, trough the PU amplifier with the normal PSU voltage setting.

My own observations:

R300 together with C313 forms a low pas filter to reduce the hum and noise in the PU stages. The -3dB point is around 3hz. By putting in a 7812 or 7815 in the PSU, this filter is no longer necessary. Keep the capacitor in place and replace R300 by a wire link. This also reduces PSU induced crosstalk. If you are consequent replace resistors R305/308 and R315/316 to maintain the DC bias current settings and the voltages as mentioned in the diagram, if you are a purist also replace all the other resistors concerning the DC bias settings of the other stages when you are applying a 7815 regulator or the 16V zener option! The 7812/7815 will not work in all situations, in some cases the input voltage of the 7812/7815 will not be sufficient. The LM2937ET-15 low voltage drop regulator is a little bit more expensive, but performs very good under all feasible conditions. So we advice to start with the LM2937ET-15.
In the diagram you will see the placement and the pin layout ( 1,2, and 3) of the regulator.

Replacing R305/308 and R315/316 is a better option than increasing R300 when you want to reduce the PSU voltage of the PU amplifier. Increasing R300 also means increasing the PSU induced crosstalk, but also a unwanted feedback effect between the two collector resistors of each channel but also between the channels. All four resistors are fed trough R300! Because the filter formed by R300 and C313 also stabilize the voltage at R305 (10,25V) and the rest, the effect will be small, but it will not be zero!
There is also a positive effect, the filter formed by R300 and C313 will have a lower -3dB point when increasing R300.
In the table the values are given for the three power supply voltages, situation one: keep R300 in place and situation two: replace R300 by a wire link.

Keep the DC PU settings:

PSU Voltage, Value in Ohms
12V, R300=560
15V, R300=2K2
16V, R300=2K7

Remove R300 (replace by wire link)!

12V, R305/308=100K, R315/316=4K4
15V, R305/308=133K, R315/316=6K5
16V, R305/308=140K, R315/316=7K2

Instead of reducing the gain of the output stages I use a resistive network to adjust the input sensitivity, like Quad themselves does concerning the tape input. Reducing the gain of the output amplifiers helps, but increasing the gain of the PU amplifier does not help. If you don’t need all the inputs, i.e. incidentally a turntable, and don’t want to connect the FM2, FM3, FM4 or AM3 tuners to the preamp then adjusting the gain of the output amplifiers and the PU amplifier is a valid option.

So to sum up: my revision advise for experienced users. Remember, there is nothing wrong with the standard upgrade, these are extra's.

If you use Quad tuners, leave the gain settings as they are. If you use other high output equipment, change the gain of the output amplifier boards. Use a LM2934ET-15 volt regulator in the PSU, remove all the unnecessary components, also the extra supply for the switching of secondary equipment. Because of the regulator, the caps in the PSU can be of relative low value, there is also the extra 100uF on the PU board. Replace R300 by a wire link in the PU amplifier and replace R305/308 and R315/316 to maintain the DC settings. Change R101/102/103/104 to 47K, this gives a more or less standard impedance of the PU input. Replace all the electrolytic capacitors. Connect a CD player to the Tape input, 1 Volt setting.

Joost Plugge

Appendix 1

The Ruud Janssen 33 upgrade and revision in a nutshell. (See the Quad-links in the left column for the original text - in Dutch). This is the basis for our standard upgrade kit.

Power Supply

Replace R500/501 (120 Ohm) by two 28 Ohm resistors. Replace MR500 by a 16V 250mW zener.
Replace C500/501 with radial capacitors of 4700uF 25V, maximum dimensions: 35 x 17mm, replace C502 with an axial type, 1000uF 25 V (30 x 16mm). Replace MR501/502 with 1N4002 diodes. Remove C503/504, R502 and MR503 (My advise).

Output amplifier boards

Reduce gain, from 5 to 2,4

Change R411/411 from 1800 Ohm to 1000 Ohm, change R412/412 from 470Ohm to 1200 Ohm

Replace:
C401/401 2,2 uF 25V
C405/405 47 uF 25V
C406/406 22 uF 25V

PU amplifier

Replace:

C300/303 47uF 25V
C301/302 100uF 25V
C307/308 47uF 25V
C311/312 47uF 25V
C313 470uF 25V

Change R300 from 560 Ohm to 1500 Ohm
(I don't agree with this value, to low!)

PU Adapter board

Replace R101/102/103/104 with 47K to change the input impedance from 68K to 47K.

Use M1 position of the adapter board to compensate for the gain reduction of the output boards. Otherwise change the values of R105/107 for the left channel, and R106/R108 for the right channel, to change the gain of the PU amp. No values are specified, only that the sum of R105/107 and R106/108 should stay the same for a proper RIAA correction. A lower value of R107 will increase the gain. A gain setting above 78x is not recommended.

Optional: replace all the resistors in the RIAA section and the first transistor stage with low noise types.

Motherboard

Replace:

C5/6 100uF 25V

18 Comments:

Anonymous Anonymous said...

I found your suggestion of improving the Quad 33 preamp very interesting. However, I have a difficult time to follow your description because I cannot find a schematic of Quad 33 which lists the part numbers. Which schematics of Quad 33 did you refer to? My schematics of Quad 33 shows only the part values, not numbers.

I studied this site as well:

http://richardbrice.net/quad33&303.htm

But, there are no part numbers.

Thanx.

7:52 PM  
Anonymous Anonymous said...

Note that the LM2937ET-15 is not a direct drop-in replacement for a 7815 although the pinout is the same. If you use an LM2937-ET15, the output capacitor (C500) needs to be at least 10uF for stability, whereas for a 7815 you don't need more than 100nf. The 7815 didn't work in mine: it couldn't muster up more than 14.5 volts output, which isn't enough for complete stability of the output voltage, or for complete hum suppression either.

Also there's a much better way to mount either of these regulator, without the wire link. Pin 3 can go as shown, pin 1 where the inboard end of the wire link is shown, and pin 2 in the hole vertically above pin 3 in the diagram, on the track that leads down to the -ve end of C502.

EJP

6:24 AM  
Anonymous Anonymous said...

Well 10uF for C500 isn't enough with the LM2937ET-15. The sound is harsh and lacks warmth & bass.

I tried 100uF which was better (less harsh) but still not right (still lacked warmth & bass), then 1000uF, which put it back to what I expect to hear from a 33.

This is now a pretty high value for the output capacitor of a regulator, which after all is supposed to be a capacitance multiplier. I suspect it's higher than necessary, and that 470uF might be enough, or even 220uF, but haven't tried them.

EJP

9:15 AM  
Anonymous Anonymous said...

Hello,

Ask Stefaan for the diagram, Esmond, thanks for the layout suggestion, I check and will change the blog.

Joost Plugge

10:20 AM  
Anonymous Anonymous said...

There are 5 outputs from the PSU board of a Quad 33. 1 and 2 go to the indicator lamp, 3 and 4 are the main +12v and 0v connections.
Pin 5 is a blue wire which links to a solder point (no other connection) marked "P" on the motherboard. Also connected to this point "P" is an Orange wire on one of the main looms which runs to the back of the pre-amp. On older production 33s this connected to Pin 4 of the output socket, which is also marked "P" on the rear fascia. On later production 33s, this wire is cut short and not connected to the output socket. What does this connection do? Why was it ever considered useful? Why did Quad change their mind about it? While writing, what voltage output are pins 1 and 2 to the lamp?
OB

11:17 PM  
Blogger Stefaan said...

This +5 Volt line was originally intended for switching on/off the power amplifier from the control-unit.

It was never developped any further until the "Quadlink" in the 6-series.

Stefaan

12:55 PM  
Anonymous Anonymous said...

Hello, I tried the 7815 from ON, used the same mod on the PCB as explained and as soon as I switched on I got a strong noise (buzzzzz) all the time. I tried a 7812 from ON that I have and the same problem occured.
I put back my old 28 years old caps and zener for now and no problem whatsoever.

I wonder if it is the "ON" regulators that are noise generator.??

8:08 PM  
Blogger Stefaan said...

Hello,

A normal 7812 needs 19V input voltage a 7815 23V, so check the voltage on pin 1. We will change the article, advising to use the LM2937ET ad first hand.

Joost Plugge

12:39 PM  
Anonymous Anonymous said...

Hello Stefaan,
Thank you for your reply.
Putting back the 7815, I found that contrairy to the previous experience, the noise was loud until it reached 15vdc, then it was much less but still there.

This is my reading, Stefaan:

Xfo sec. = 20vac
Vdc unregulated (pin 1) = 19.95v
Vdc regulated (pin 3) = 14.99v

As the V drop for the 7800 series is 2v so I thought I had enough with an input of 19.9 for Vin.

I can try the LM2937ET but:
1- What about the LM317 adjusted for 15v, apparently they are more quiet than the 1800 series, would they accept a Vin of 19vdc?

2- What about if we use the 6vac of the other secondary (the unused sec), disconecting it from the ground, add a full bridge of 4 diodes and add it in serie with the Vin 19vdc (connect the 0v of the 6vdc to the +19vdc) which would give a 24 to 26vdc approx.? Is that unused xfo sec built strong enough to take the little juice?
I tried to make a drawing below.

Xfo
sec diodes

(--+---|>|--------------+--o+6vdc
( | |
( `---|<|---. |
( | |
( ,---|<|---+--o 0vdc |
( | |
(--+---|>|--------------'

(-----|>|---+--o+19vdc
( |
(-----o Gnd |
( |
(-----|>|---'
Connect 0vdc with Gnd

This is a question and suggestion.

Thank you

9:10 PM  
Anonymous Anonymous said...

I'm afraid my drawing doesn't look good as it didn't come out the way I typed it. Sorry!

11:30 PM  
Blogger Stefaan said...

Hello,

The dropout voltage of a 78XX is 2V, for the LM2937 this is 0.5V. So the input voltage could be to low with the 78XX, because you also have to take the ripple in to account, The LM will work, also the 7812 should work. What is the value of the input capacitor? I never tried to connect all the windings in series, but it is a interesting solution. When you are using a bridge rectifier, this should work also, keep us posted.

Joost Plugge

8:00 PM  
Anonymous Anonymous said...

Thank you to remind me that there is still ripple and it is rejected by the regulator. My cap is a Panasonic 2200uF 35v.

I don't know what happened though that the noise was still there with the 7812 only not quite as loud as with the 7815, and it was permanent as now after reinstalling the 7815 for test the noise decreases a lot when the 15v is reached so the 7812 should be ok now. My question now is what could create such a diffrence between the two test of the 7815 and using the same components in both test? I personaly don't have the answer.
Daniel

3:00 PM  
Blogger Stefaan said...

Hello Daniel,
I checked my own unit, the input voltage of the regulator was 19V, so the LM worked ok, as a 7815 would/should do. The mains input was 230 VAC. With the input capacitor you mentioned, the ripple should be very low. If the output capacitor is 10uF or higher, as EJP sugested, it should work fine. I can't explain the behauvior of your PSU, you need a osciloscoop to check it for paterns en the frequency of the signals.
Regards Joost Plugge

9:15 AM  
Anonymous Anonymous said...

Thanks Joost,
Unfortunatly I only have a meter but I'll make experiment to find by ear what is the cause or at least how I stopped it if I succeed.
I'll do that before using other alternative and I'll give you the result so that it could help others who might have the same conditions as me.
I'm in Quebec, Canada, the main here is 115-125v 60Hz.

Thanks again

Daniel

5:06 PM  
Anonymous Anonymous said...

Well about this hum, the only way I eliminated it is by keeping the 680uF (or 470uF) with the 120 ohms before the 2200uF 120 ohm and 2200uF just like the original design before the regulator but the Vin before the regulator gets a bit low for the 7815, I still get 14.90v as a Vout and no noise. If I remove any of the two 120 ohms siply the 680uF, hum is back. So to conclude, I just changed the zener for the 7815 and changed the two 1000uF for two 2200uF and to my earing evering thing is fine, as soon I have the low drop out regulator then I will replace the 7815 to make sure I have enough room for a good regulation as for now the Vin - Vout is nearly less than my the drop out of the 7815.

Daniel

10:30 PM  
Anonymous Anonymous said...

Hello,
I also found that I had a bad contact at the ground pin of the output din plug where my Quad 405-1 is connected which causes a hum and desapears as I move the plug. Now no more hum, every thing is fine.

Daniel

5:11 PM  
Anonymous Anonymous said...

Hello,

I made the modification trying either with a 7815 or a LM2937. The result is absolutely great. But, there's two points of view on which the result isn't as good as before :
(a) firstly I have a little hum (with both regulators) even trying varied capacitors at C500 (100nf, 10uf, 100uf, 2200uf).
(b) secondly the murmur is much more audible than before, I can here it from where I listen the music when there's no music. I tried putting a 100nf capacitor in parallel with C500, it isn't really better. Sadly, I can't manage to know what is the cause. Perhaps it isn't the regulator, but I'm pretty sure it is.

3:33 PM  
Anonymous EJP said...

The 10uF capacitor should not be placed as shown. It should be between the diodes and the 7812, where the zener was, with the same polarity as the diodes. Otherwise it is off on a tee, which is frowned on in PSUs.

5:35 AM  

Post a Comment

<< Home