The only advantage would be if you intend to use a supply voltage over +/-45V. Under 45V, BC550C/BC560C are likely to be a better choice.
In order of priority:
- The 2.2K feedback resistors should be matched as close as possible. Usually, from about ten resistors, you can get two pairs that are within 0.1~0.2% of each other, fairly easily.
- The 100R resistors, R7 and R8. Along with the 2.2K those set the gain.
- You can try to match the input transistors, the BC550/560 (or BC546/556) pairs.
- VAS transistors (yours will be close already)
Note that according to Lazy Cat, the transistors just should be closer than 15%.
Terry,
The main reason one would go with the BC546C / BC556C would be for higher voltages, since they are 65V rated parts. These do have a higher 'Noise Figure' than the BC550C / BC560C, though I doubt it will make any meaningful difference. Also note that the BC556C is an obsolete part, where the BC550C / BC560C are current production.
Most of the parameters within this family of parts are the same aside from those already mentioned. For 40V rails or less you would do well with the BC550C / BC560C. That way in the event of an 'oops' you aren't sacrificing NOS Unobtanium to the silicon gods
.
Match the feedback resistors as best you can and shoot for a reasonable match in the input transistors and VAS transistors, though these are not super critical.
D'oh! typing too slow. Pete beat me to it.
The main reason one would go with the BC546C / BC556C would be for higher voltages, since they are 65V rated parts. These do have a higher 'Noise Figure' than the BC550C / BC560C, though I doubt it will make any meaningful difference. Also note that the BC556C is an obsolete part, where the BC550C / BC560C are current production.
Most of the parameters within this family of parts are the same aside from those already mentioned. For 40V rails or less you would do well with the BC550C / BC560C. That way in the event of an 'oops' you aren't sacrificing NOS Unobtanium to the silicon gods
.Match the feedback resistors as best you can and shoot for a reasonable match in the input transistors and VAS transistors, though these are not super critical.
D'oh! typing too slow. Pete beat me to it.
Last edited:
I managed to get more than enough for anyone who wants to use these, from another member. Same thing applies to the KSC3503, "E" version.
For now, BC550C/BC560C and KSA1381E/KSC3503E will make the builds close to Lazy Cat's original component selection, and remove any uncertainty. Having said that, there is no reason not to experiment with different parts... as far as I am concerned, that makes DIY projects fun.
Eventually, we will have alternatives, thanks you your efforts, and to other people who have taken an interest in the component selection. Tweaking the layout for a different TO-92 pinout should not stop the show, either...
edit:
I sent out a few sets of the BC546C/556C... to you, because I know you are working on comparing the different options, to Terry, because I knew he was interested in the higher voltage options already, and a few other people so they would have them on hand if needed. It was not because I think they are better than the more common BC550C/560C.
Last edited:
I have, yes. Good to hear from you, and I hope all is well. I was a bit worried.
I will post more information on components, but the short story is this:
I have enough of the currently hard to get parts, for everyone who has expressed interest in the boards so far, and probably lots of spares. This does not mean you have to use those exact parts for a good build.
Shaan has shown that you can get very good sound from this circuit with commonly available transistors, like the BD139/140.
The point of having the KSA1381-E/KSC3503-E, is that we can build the boards with parts which are close to what Lazy cat specified for the original VSSA module, IF we want to.
The reason for the BC546C/556C is higher supply voltage (above 45V), or to allow for a little headroom for line and load regulation. Below +/-45V, the BC550C/BC560C is just as good, or an even better choice.
I have enough of the currently hard to get parts, for everyone who has expressed interest in the boards so far, and probably lots of spares. This does not mean you have to use those exact parts for a good build.
Shaan has shown that you can get very good sound from this circuit with commonly available transistors, like the BD139/140.
The point of having the KSA1381-E/KSC3503-E, is that we can build the boards with parts which are close to what Lazy cat specified for the original VSSA module, IF we want to.
The reason for the BC546C/556C is higher supply voltage (above 45V), or to allow for a little headroom for line and load regulation. Below +/-45V, the BC550C/BC560C is just as good, or an even better choice.
Well I have the boards populated. I used BC546/556 to give myself the option of going higher rails if I buy a bigger transformer. Will see if there is a noticeable difference in sound with these trannies.
I don't have any 20-22r for the zobel. Can I use 10R there?
Also, what is a good starting point for the 2K pots?
Thanks, Terry
I don't have any 20-22r for the zobel. Can I use 10R there?
Also, what is a good starting point for the 2K pots?
Thanks, Terry
That was quick!
You can use a single cap and a single resistor, In other words, 0.1u in series with 10R, but then the 10R should be at least 1/2W, and ideally 1W. You can also just leave it off, for testing, especially if your speaker wires are fairly short.
I set the pots mid-way to start, they usually come that way from the factory, anyway.
A better way, is if you adjust the pots after assembly, and before first power-up, so that the resistance across the pot reads approximately 1.5 K Ohms, close to the calibration setting. It may seem like a pain at first, but it makes calibration easier.
You can use the two empty holes adjacent to the pot as test points.
Attachments
I would adjust to 12~15 mA VAS bias current and 0V +/-5 mV offset, at least in the beginning. After that, you can change it if you think it is too low or too high in your case. I will just paste my preliminary calibration procedure into a post below.
Offset and Bias Adjustment
Overview:
The two trim pots are complementary. For best results, they should always BOTH be adjusted the SAME number of turns or same fraction of a turn. (you do not have to follow this suggestion to get a result, but it may take several iterations otherwise).
The trim pots can be pre-set to 1.5K Ohms, measured across the outside trim pot terminals on the bottom of the board (or from R5 to R7 on top) before applying power. Otherwise, leave them at mid-travel.
The output DC offset voltage can be measured directly. The bias current will be measured indirectly, as a voltage across a 10-ohm resistor (R9, R10), using TP1 or TP2.
To make the subsequent bias adjustment easier, set the output offset voltage to approximately zero first, then set the bias current to the desired level, and make any fine adjustments to the offset at the end of the procedure.
If needed, the input bias current can be checked by measuring the voltage across R5 and R6, and dividing by the resistor value. It should be about 1.9 mA.
Getting Started:
Before applying power, the board should be installed on a heatsink. It will dissipate about 10W during calibration.
20~22 Ohm 5~10W resistors should be installed in series with the positive and negative power supply wires. You can also use the popular "Bulb Tester" method, incandescent lamps (bulbs) in series with the supply rails.
A small jumper should be installed at the input terminals to guarantee zero DC input voltage.
After applying power for the first time, check the supply voltage at the positive and negative terminal on the circuit board, or at D3 and D4. The supply voltage should be approximately what you expected and the voltage drop across the 22R should be less than 5V.
Calibration Example:
Note:
This is a slow and fairly safe way to adjust the offset and bias, and requires only one meter to be used (it may not be the quickest, or the most accurate). There are 5~6 alternative methods already posted in the PeeCeeBee thread and the main VSSA thread.
Anyone who wants to, should feel free to post a favorite approach below, or help others learn a better way...
Overview:
The two trim pots are complementary. For best results, they should always BOTH be adjusted the SAME number of turns or same fraction of a turn. (you do not have to follow this suggestion to get a result, but it may take several iterations otherwise).
The trim pots can be pre-set to 1.5K Ohms, measured across the outside trim pot terminals on the bottom of the board (or from R5 to R7 on top) before applying power. Otherwise, leave them at mid-travel.
The output DC offset voltage can be measured directly. The bias current will be measured indirectly, as a voltage across a 10-ohm resistor (R9, R10), using TP1 or TP2.
To make the subsequent bias adjustment easier, set the output offset voltage to approximately zero first, then set the bias current to the desired level, and make any fine adjustments to the offset at the end of the procedure.
If needed, the input bias current can be checked by measuring the voltage across R5 and R6, and dividing by the resistor value. It should be about 1.9 mA.
Getting Started:
Before applying power, the board should be installed on a heatsink. It will dissipate about 10W during calibration.
20~22 Ohm 5~10W resistors should be installed in series with the positive and negative power supply wires. You can also use the popular "Bulb Tester" method, incandescent lamps (bulbs) in series with the supply rails.
A small jumper should be installed at the input terminals to guarantee zero DC input voltage.
After applying power for the first time, check the supply voltage at the positive and negative terminal on the circuit board, or at D3 and D4. The supply voltage should be approximately what you expected and the voltage drop across the 22R should be less than 5V.
Calibration Example:
- To adjust Offset Voltage, connect a voltmeter from the power supply ground terminal to the output terminal. The meter measures the DC offset voltage directly at the output. Adjust BOTH trim pots clockwise to lower, and counter-clockwise to raise the offset voltage, until the meter reads 0 +/-5 mV.
- To adjust Bias Current, connect a voltmeter from the inside lead of R11 to TP1. The meter measures voltage across a 10 ohm resistor. Adjust BOTH trim pots in opposite directions, until the meter reads 140 mV (0.140V). This will correspond to 14 mA VAS bias current. At this point, input bias should be about 1.9 mA. Assuming a supply voltage of about 35V, total idle current should be about 150 mA.
- Quiescent power will be about 10.5 watts per channel - this is why the board must be attached to a heatsink any time power is applied (!)
- Offset and Bias should both be checked, and fine adjustments made, when the heatsink temperature has stabilized.
Note:
This is a slow and fairly safe way to adjust the offset and bias, and requires only one meter to be used (it may not be the quickest, or the most accurate). There are 5~6 alternative methods already posted in the PeeCeeBee thread and the main VSSA thread.
Anyone who wants to, should feel free to post a favorite approach below, or help others learn a better way...
Attachments
Last edited:
I am not sure if it will make a big difference, but detecting a difference, or measuring it, may be difficult.Pete,
Thanks for the write up's. I will be starting my builds soon. Here are my 1st 2 DIY Boards. Red one is .032" 1/2 oz Copper. Other is .064" 1 oz Copper. Looking to see if Copper thickness makes any difference.
Rick
Both boards look great! Did both etch about equally well?
Thanks Pete, that is exactly what I needed. I set the VAS at 140mv and that took the overall bias to 152mA measured at the + rail with an ammeter. I used two meters to set the bias and it worked really well. Takes a minute to get used to twisting two screwdrivers at the same time but once I got the hang of it, things went pretty well. I have had it playing for about an hour now so I could be sure the heat sinks would be sufficient. Looking good so far. In a while I plan to hook up Jason's TO-3 design and A/B them to see if I can hear any difference.
Blessings, Terry
Blessings, Terry
Attachments
You don't actually have to do it simultaneously. I usually do them alternating, one turn on the left, then one on the right, etc.
Congratulations, you are officially the first one besides me to get the new boards playing music!
According to Lazy Cat, it takes up to 50 hours (!!) before the electrolytics settle down. I can't verify that long, but when I first assembled the second board, after listening to the first one for a few days, they sounded different. (I swapped speaker channels, just to make sure). About a 1-2 days later, they were the same, to my ears, anyway.
I am curious, and also for anyone else - how long did it take you to calibrate the boards?
edit:
Do you have VAS compensation caps installed on the boards?
Last edited:
Hi Pete,
The first board took probably upwards of 10 minutes to set because I first set the offset and then the VAS. While doing that I realized that the offset changes as you turn the pots to set the bias. Then when you readjust the offset it changes the bias. Felt like I was chasing my tail. Once I hooked up two meters I was able to keep the offset low while I was adjusting the VAS so when I got to around 140mV it only took a few slight adjustment to get everything settled. The second board took less than 5 minutes including hooking everything up.
I'm running the rails at about +/-37.8vDC. I hooked it up to my scope and it begins to clip at about 22vAC into an 8ohm dummy load.
I have both yours and Jason's playing side by side and hooked up to an A/B switch. They are very close in sound but not exactly. There is just a slightly different voicing between the two. You have to listen closely to hear it. I will assume it is mostly the caps breaking in. Jason's boards have a couple days play time and of course yours are spanking new. Still fun comparing.
Yes, I installed 47pf mica in C11,12.
Blessings, Terry
Edit; I forgot to mention, I installed the VAS on the main heatsink. Mainly because I didn't have any of those cool tiny heatsinks you used. My question is how did you plan to attach them to the heatsink? I had to drill out the holes in the PCB to be able to fit the screw heads through to the top of the transistors. The MOSFETS are tight against the bottom of the PCB so I just tightened the PCB down onto the top of them but because the VAS trannies are so much thinner you can't do that with them. Just something to think about.
The first board took probably upwards of 10 minutes to set because I first set the offset and then the VAS. While doing that I realized that the offset changes as you turn the pots to set the bias. Then when you readjust the offset it changes the bias. Felt like I was chasing my tail. Once I hooked up two meters I was able to keep the offset low while I was adjusting the VAS so when I got to around 140mV it only took a few slight adjustment to get everything settled. The second board took less than 5 minutes including hooking everything up.
I'm running the rails at about +/-37.8vDC. I hooked it up to my scope and it begins to clip at about 22vAC into an 8ohm dummy load.
I have both yours and Jason's playing side by side and hooked up to an A/B switch. They are very close in sound but not exactly. There is just a slightly different voicing between the two. You have to listen closely to hear it. I will assume it is mostly the caps breaking in. Jason's boards have a couple days play time and of course yours are spanking new. Still fun comparing.
Yes, I installed 47pf mica in C11,12.
Blessings, Terry
Edit; I forgot to mention, I installed the VAS on the main heatsink. Mainly because I didn't have any of those cool tiny heatsinks you used. My question is how did you plan to attach them to the heatsink? I had to drill out the holes in the PCB to be able to fit the screw heads through to the top of the transistors. The MOSFETS are tight against the bottom of the PCB so I just tightened the PCB down onto the top of them but because the VAS trannies are so much thinner you can't do that with them. Just something to think about.
Attachments
Last edited: