Salas hotrodded blue DCB1 build

School

Pleas explain the purpose of the LED's.

Are they used as a resistor?

If not matched, what is the consequence?
 

Attachments

  • CCTblue.gif
    CCTblue.gif
    28.7 KB · Views: 4,681
Received first of 3 boards this morning. Lovely! Many thanks.

I am going to use it for RIAA supply at 27v. I will only need the + rail. How do I isolate the other rail regarding the diodes and power supply caps? Will I need only 2xMUR820, or is it still as I assume x4 I need? Presumably I only need x2 of the x4 caps. Will I need to add a jumper somewhere to link it up?

What transformer is recommended for a 27v single rail device with this reg please?

Many thanks
Lucas
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Received first of 3 boards this morning. Lovely! Many thanks.

I am going to use it for RIAA supply at 27v. I will only need the + rail. How do I isolate the other rail regarding the diodes and power supply caps? Will I need only 2xMUR820, or is it still as I assume x4 I need? Presumably I only need x2 of the x4 caps. Will I need to add a jumper somewhere to link it up?

What transformer is recommended for a 27v single rail device with this reg please?

Many thanks
Lucas

There, full guide and users applications. Also for your questions after reading at least post#1. http://www.diyaudio.com/forums/power-supplies/153484-using-hypnotize-general-shunt-reg-pcb.html
 
there was an earlier poster who claimed that he got better sound by running the amplifying FET at a bias above it's Idss.

I disagreed then and still do now.
A jFET follower will tolerate some higher current running during signal induced high Vgs, read Curl. But there is a big downside to using a high bias at above Idss and then applying the signal induced Vgs sending current even higher.

I would much rather run the bias at 95% of Idss rather than set bias to 105% of Idss
What the B1, as proposed by Pass, does is run the FET follower at very near but not above 100% Idss.

Ah, so have I misunderstood? I thought that I had read that the conventional arrangement is for the higher Ids to serve as the CCS FET.
Perhaps I missed that in the original B1 guide. I did not do my own matching on the FET's but bought a few sets from a member source. Since I had 2 packages marked as 7.7 Ids I used those.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
If the audio quad JFETs are very closely matched per pair its good for offset, and I recommend best match possible especially in the Blue. If tight, an average DVM will not be dependable to show 2 times which one is slightly less Idss anyway, and if you blow over or touch they will jitter. But if they have a slight and reliable difference, use the lesser one for inner position per pair.
 
There, full guide and users applications. Also for your questions after reading at least post#1. http://www.diyaudio.com/forums/power-supplies/153484-using-hypnotize-general-shunt-reg-pcb.html

Hi Salas,

Sorry to be a pest, but I am reading and reading, but I can't find advice on specifics relating to single rail set-up anywhere. I realise that the central bar is a common earth, and that I need populate the + side only, but it's not so simple at the PS caps and rectifier diodes, which span across the two sides. Do I need to add a jumper from the - side to earth to complete the rectifier circuit?

Many thanks
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Did you understand the setting procedure and saw a single rail T-amp project there? If yes, up to here all well, then please post your above rest of questions as new post in that thread. We will discuss it there soon, all the general reg things are there. Thanx.
 
Whats the procedure to match the LED's

Matching the LED'S, I wire a 1.2K resistor in series with the LED apply 9vdc and measure the voltage across the LED?

Vf what's the f stand for voltage what ?

What voltage am I looking for? I ordered 40LED's and need
16, so I measure all 40 and select the 16 that measure the same
or as close to the same?

Please advise.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Matching the LED'S, I wire a 1.2K resistor in series with the LED apply 9vdc and measure the voltage across the LED?

Vf what's the f stand for voltage what ?

What voltage am I looking for? I ordered 40LED's and need
16, so I measure all 40 and select the 16 that measure the same
or as close to the same?

Please advise.

Yes you can do that.

Vf= Forward voltage drop. Describes the nominal voltage that any diode develops when sufficient current is run through it given its nominal spec. LED is a glowing diode you can say.

You look for enough measuring as close to the same between them.
 
Member
Joined 2009
Paid Member
Pleas explain the purpose of the LED's.

Are they used as a resistor?

If not matched, what is the consequence?

They are the active low noise voltage reference. If not matched, small assymetry in +/- Vout.

Are there any sources out there to get matched LED´s?
Not that it would be difficult, but i think not everyone want to buy a lot of LED´s to get 10 matched, me including. :eek:

Perhaps someone would like to do the job and make a GB?
 
Are there any sources out there to get matched LED´s?
Not that it would be difficult, but i think not everyone want to buy a lot of LED´s to get 10 matched, me including. :eek:

Perhaps someone would like to do the job and make a GB?

I think the key issue is to get the STRINGS voltage matched. Having all the individual LEDs voltage matched is not necessary. This gives you a quite a bit of leeway in LED selection. :)
 
I think the key issue is to get the STRINGS voltage matched. Having all the individual LEDs voltage matched is not necessary. This gives you a quite a bit of leeway in LED selection. :)

What does that mean? "Strings voltage match"

I am still trying to understand this circuit, using LED's to "they are the active low noise voltage reference"

LED's have no opposition to current flow so having several LED's in series does what to the voltage? what is the voltage going in vs what is coming out

or is the voltage the same only a more stable voltage? How would a reference voltage be achieved if no LED's were used? What circuit would be used?
 
I think the key issue is to get the STRINGS voltage matched. Having all the individual LEDs voltage matched is not necessary. This gives you a quite a bit of leeway in LED selection. :)
well said Mach1.

The LEDs don't need to be matched.
It is helpful to measure the Vf and label them accordingly, but not necessary.

A batch of LEDs will be very close in Vf. I have 5000 off of red LEDs that are 1.55Vf A tolerance of +-0.05V will probably bring 99% of them within 1.5V to 1.599V
I also have 100 red LEDs that are 1.95Vf+-0.05V. This batch are obviously made by a different manufacturing process. I have another batch that are ~1.8Vf

Selecting from these bags allows an almost infinite combination of string voltages.

The LEDs in the DCB1 are in three groups. A singleton, a triplet and a quintuplet.

The singleton is a confirmation that power is available after the 12V regulator.
It can be any LED, red, green, blue, white, orange, yellow.
The triplet is a string of three unmatched LEDs that have a total Vf between 5V and 6V. If the strings of total voltage are made consistent for all your Shunt regulators, it makes for less calculation in setting the current in the CCS part of the regulator. Again it does not matter if the strings are different, but it saves time and measurement and adjustment later if you know beforehand that your strings all measure 5.85V to 5.9V or any other range you come up with.

The purpose of this triplet is to set the Vgs of the CCS mosFET. The Vgs is a very variable parameter of mosFETs. All that time you spent matching your triplet string is thrown out the window because the mosFET has a much wider Vgs tolerance than the LEDs.
Best to just select a random mosFET and a group of three random LEDs and assemble with just one of the parallel resistors fitted. Measure the voltage drop across the CCS resistor. What is the current? How much more current do you want? What value of resistor will you select to go into the parallel location. Job done.

Finally the quintuplet.
This sets the output voltage of the shunt regulator.
The output voltage is the 5LED string Vf plus the Vbe of the adjacent BJT plus the voltage drop of the resistor between LEDs and BJT base.
The resistor Vdrop is very low due to the tiny base current flowing through it. Ignore this.
The Vbe of the BJT will depend on the type of device chosen. Choose a random device from the bag of recommended devices. It's Vbe will be between 550mV and 700mV but it's tolerance will be within about +-10mV for that batch of devices. Lets suppose it's 585mVbe +-10mV.
The string of 5LEDs had a total Vf of ~8.9V
The output voltage will be ~8.9+0.585V +-10mV. Vout = 9.485V+-10mV

Compare this to the other half of the DCB1. Swap a few LEDs in either 5LED string if you feel that the imbalance in regulated voltage is outside your desired tolerance.

IMPORTANT
The tolerance of the CCS current does not affect the regulated output voltage.
 
Last edited:
Member
Joined 2009
Paid Member
When I build my DCB1 to make things "safe" I will need to install a capacitor on the input of my yet to be built F5 [still waiting for the heatsinks; 6 weeks & counting] What value cap do people recommend and would it make any difference if I installed one on the F5 input or the output of the DCB1? I am reading through all relevant threads so hope I havent missed this bit of info.
Thanks.http://www.diyaudio.com/forums/images/smilies/confused.gif
 
The DCB1 is by definition DC coupled.
The F5 is also DC coupled.

When both the source and the amplifier have no DC blocking capacitor, I believe it is essential to protect the speaker from abuse.

I recommend that a DC detect system be added. On detection of excessive a system to protect the speaker must be added.
I would also recommend, but not essential, that a DC servo be added to minimise the drift in output offset.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
*If someone is perfectionist about matching the Verf LED gangs of 5, the best maniac party trick is to use a *12V battery, connect gate and source together in the specific 2SK170s that are going to feed those strings (those Jfets above 1R), feed (+) to the drain, connect g&s node to LEDS string, return string's other end to (-), and swap between enough glow bugs until you got 2 total across 5 LEDS Voltage matches, each to its own Jfet. For the cherry on the cake set DVM that measures diode drop (some got it) and find 2 Vbe matched BC550 & 560.
Practical benefit? Next to nothing. Aesthetic ritualistic benefit? Priceless.:D

*9V,1.5V,1.5V batteries in series that are common types, maybe already in remote controls on the coffee table, will do for 12V total. And don't post pics from the parking space tapping on the car's stuff.:)
 

Attachments

  • MatchedInHeaven.gif
    MatchedInHeaven.gif
    2.6 KB · Views: 3,768
Member
Joined 2003
Paid Member
When I build my DCB1 to make things "safe" I will need to install a capacitor on the input of my yet to be built F5 [still waiting for the heatsinks; 6 weeks & counting] What value cap do people recommend and would it make any difference if I installed one on the F5 input or the output of the DCB1? I am reading through all relevant threads so hope I havent missed this bit of info.
Thanks.http://www.diyaudio.com/forums/images/smilies/confused.gif

I've been asking around about that too. The suggestions I've had is to put it on the input of the F-5. I would prefer to leave the F-5 as is, and put it on the output of the B-1, although it wont be exactly DC then, which was the point, but I like a little insurance. The person who was commenting seemed to suggest it would be better on the input of the F-5, I'm not clear on why, so I continue to ask around. Nelson apparently chose to put the caps in the B-1 (standard non DC version of course) for what ever reason.

Russellc