attempt to convert B1 into DCB1

[spirovious]

Hi,
I read much good things about DCB1-
http://www.hifivision.com/diy/53964-dcb1-build.html

Got chance to demo and compared it with B1 at FM's places.Yes I found DCB1 to be better.Since it does not use coupling caps(DC-Coupled B1),it is said to sound better.I started reading and found that a simple DCB1 can be built by using existing B1 parts on separate raw PCB.

Salas DCB1 circuit-

Original B1 circuit-

Checked DCB1 parts and found it uses -

Resistors-
220R X 2,
220K X 1
1M X 1
JFET same as B1 X 2

I removed parts from one channel of B1 and mounted on raw PCB-

\

I double checked circuit before powering it on.
Ran Right channel in DCB1 and Left in original B1

First impression-
Mids and details are much better than B1. Sound is open and pleasant with dark backgound.
Extended bass in some songs disappeared.Marginally DCB1 is louder than B1.

Now I would like experts to comment if I have done something wrong in circuit.I hope this simple connection does not harm JFETs or poweramp.

Thanks.

 

[manniraj]

@spiro, I think their is a voltage difference that you need to set on the PSU for DCB1 (reduce).

 

[spirovious]

@spiro, I think their is a voltage difference that you need to set on the PSU for DCB1 (reduce).

Yes,I used 10-0-10V from PS for dcb1 and around 20V for B1.

Resistor values-

B1 has 221R instead of 220R, so used the same(unlike in figure).

 

[omishra]

1M need ground reference.
This may ssound shrill. Needs de coupling caps. Also star ground reduces HF noise.

 

[spirovious]

1M need ground reference.
This may ssound shrill. Needs de coupling caps. Also star ground reduces HF noise.

Thanks.l simply followed Salas design.
He has not used any caps.May be his powersupply is designed considering it.Now our b1ps also has many caps.So do we really need caps if no audible noise is traced?
Any issue running poweramp without using caps?

 

[omishra]

Salas is using shunt regulator. It has low output impedance. We use series linear regulator with resistances and coils in path. It needs capacitors to lower impedance. I am using 15000uF. But minimum 100uF required with correct polarity near each jfet. BASS and softness of HF is heavily depend on this.

Regarding usage, i am using it into TV setup as well as pass F5 turbo. My Pass Amp setup is E2E capacitor less chain. Schiit biforst Uber also has capacitor less output coupling. Almost 1 year no problem so far. I know certainly i am not playing with live wires anytime.

 

[spirovious]

I am using 15000uF. But minimum 100uF required with correct polarity near each jfet. BASS and softness of HF is heavily depend on this.

I can see them in design-

Few queries in above image-

1.Above Q1,there is C1.So is it connected to PS ground(0v)?So that capacitor lies between +10V and Ground(zero V).
2.Also C2 next to Q2 connected to ground only? That is -10Volts and ground(0v)?
3.Finally is C3 parallel to C1?

 

[omishra]

I can see them in design-

Few queries in above image-

1.Above Q1,there is C1.So is it connected to PS ground(0v)?So that capacitor lies between +10V and Ground(zero V).
2.Also C2 next to Q2 connected to ground only? That is -10Volts and ground(0v)?
3.Finally is C3 parallel to C1?

Yes all true. Big cap paralleled with small film cap.

 

[spirovious]

Yes all true. Big cap paralleled with small film cap.

Thanks,any perticular reason for using different caps in parallel than to use single with same value?
Another thing,output Resistor(impedance) in earlier B1 was 1K,later used 221 ohms. DCB1 uses 220 as output resistor.So will there be any difference in sound if its value is changed like modded B1? Any risk ?

 

[omishra]

Small cap parallel to big electrolyte cap reduces inductance effect from bigger cap. Here its non relevant but suspicious mind also i wanted each jfet its own current buffer. Noise shunted also gone to sink with separate track beneath.
Resistance is there to stop oscillation as well as protect devices from small DC. No other effect on sound. At least i never paid attention, because i never noticed anything lacking so far.

 

[mbhangui]

Thanks,any perticular reason for using different caps in parallel than to use single with same value?

Electrolytic capacitors have high ESR - https://en.wikipedia.org/wiki/Equivalent_series_resistance.

The objective of a capacitor is to rapidly respond to transient. Give power when voltage drops and take power to charge when voltage rises. High ESR slows down this ability. Ceramic capacitors have low ESR. Combining a ceramic capacitor in parallel reduces the effective ESR. The smaller caps are called bypass caps, but their main purpose is to deal with high frequency spikes. They have to be small to discharge and charge quickly in response to how often the spikes come in.

At high frequencies even the traces on pcb, the length of the leads of components behave like inductors and provide resistance. That's why you see in high speed op-amps, the capacitor should be soldered to the op amp pins with as short leads as possible.

 

[omishra]

Just to add - spikes are not only generated while charging but its could be discharging also. So effect on HF is for sure.

 

[spirovious]

Finally I could complete it today. :yahoo:
Thanks to Om who helped me to workout some errors.

Tried smallest signal path

First Impression-
Very open sound.Output is true to input.I had always problem with B1 that it used to reduce output and frequency response,soundstage used to depend of output resistor.Here its not the case with DCB1. Output level matches the input one(more than B1).Also DAC sound quality is almost maintained in output.
Bass is fast,mids very clear.Treble not harsh.
I can say improvement over modded b1 is 20% or even more.
Need some time to run.Also caps used are local and ceramic caps yet to add as suggested.

 

[omishra]

Thanks for kind words, i hope your electrolyte caps have correct polarity.

 

[spirovious]

Thanks for kind words, i hope your electrolyte caps have correct polarity.

Yes dear friend You can make out from top that -ve marks are inside of each channel.Double checked circuit before running it.

 

[Hari Iyer]

@ spiro.
Great diy build quality. You have to take care that none of the power supply + -10v or +10v fails. If either of them fails the output will show full DC 10v which will be amplified and will short the woofer coils due to high current flow through its coils and get burnt.

In-order to avoid this you can either have a DC blocking capacitor or a DC detect circuit. If FMs are worried about the having the DC blocking capacitor which can hamper with the signal quality then you can have an easy DC detect circuit which can be operated with a simple 5 v relay. I will post this simple ckt today post lunch after i draw them. The DC detect ckt will detect this psu failure and will isolate the output stage immediately there by protecting the amp and the speakers in case of absence of dc blocking capacitor.

 

[Hari Iyer]

Here goes the simple ckt diag for the DC detect and protector circuit.

You can power them from a separate 5 V power supply using 7805 power regulator. If you use the same power supply used in your DCB1 then there is a risk of it failing incase they themself fail.

 

[omishra]

Here goes the simple ckt diag for the DC detect and protector circuit.

You can power them from a separate 5 V power supply using 7805 power regulator. If you use the same power supply used in your DCB1 then there is a risk of it failing incase they themself fail.

Why not have simple 24V DPDT relay operated directly from +10,-10 and pass it to load? It will cutoff ASAP either rail fails and stops power to DCB1 totally. Don't forget diode across relay coil.
I will see your circuit when i reach at home.
Edit: just saw it Hari.
It wont detect -ve voltage.

 

[Hari Iyer]

Maybe you can clone with pnp transistors too with the same input for detecting the -ve supply rail failure. using transistors will detect as low as 0.7volt DC to be detected and activate the relay for protection.

Disclaimer: I have not tested this anyway. Just a ckt idea. Will incorporate when i design my DCB1 pcb.

 

[spirovious]

@ spiro.
You have to take care that none of the power supply + -10v or +10v fails. If either of them fails the output will show full DC 10v which will be amplified and will short the woofer coils due to high current flow through its coils and get burnt.

So much things to consider . Hope PS never falls else so much worries.:sad:
Will that be issue with SSP too?

Can I use Russian Caps which came with B1 in output stage?

One doubt- In DCB1 circuit,the second JFET do not handle audio signal as the first I suppose. Isn't that helping to ground DC voltage coming from first JFET as its both pins are ground ? Please let me know.