One more tweaked V-CAN
Jul 8, 2017 11:14:03 GMT
Post by Pengo on Jul 8, 2017 11:14:03 GMT
I am a new guy here, but read some threads before. This is my V-CAN that I purchased second-hand in Russia for just 38£ (in rubles certainly 
). God knows why the owner got rid of it for this amount of money. It was in mint stock condition.
Since I'm a bit of an engineer, I decided to get into it with a soldering iron. Board routing is nice, no SMD components, it is very easy to modify and repair the board.
Stock solder, as far as understand, is trendy leadless substance. It is sh*t that is prohibited for use in hirel products in Russia. I want my amplifier to be reliable, that is why I use old-fashioned russian POS-61 (61% Sn / 39% Pb) or WBT-0800. This amp was soldered with WBT. I will not throw out it so I do not worry for the nature.
Power protection diode still in place, 2х 1000 uF caps are changed to Panasonic FR 1500 uF 25 V, FR 470 uF 25 V near them. Another FR 1500 uF 25 V near opamp. These caps are rather small, cheap, has low ESR and high endurance. That is why I use them in power circuits.
Then 6x 220 uF 16 V Nichicon FG near opamp and 2x 10 uF 50 V Nichicon FG. Probably not the best ones, but better than stock. Ceramic caps are changed to 220 pf WIMAs, 2x matched 5,1 KOhm russian MLT (metal film) gain resistors.
Input caps - Nichicon ES 4.7 uF 50 V, output caps - Nichicon ES 470 uF 35 V. Amp drives 250 Ohm DT770Pro headphones, thus 470 is enough regarding proper cutoff frequency. It would be better to use 1000 uF 16 V at the output, but I didn't find them in stock in Moscow, and my impatience said "use 470" Electrolytic caps at the input is not the best solution, but I do not want to omit them. There is not enough space to place PP foil caps, maybe I will try PS/PET coated caps later.
Output stage is ok, no mods here.
Now move on to the next part - opamp. ST MC33079 in stock. According to the datasheet:
■ Voltage noise: 4.5 nV/√Hz
■ Gain bandwidth product: 15 MHz
■ Slew rate: 7 V/μs
■ Distortion: 0.002%
Not so bad, frankly speaking. But it was changed for IC holder. I tried OPA404 (on pics), it's nice old amp:
■ Gain bandwidth product: 6.4 MHz
■ Slew rate: 35V/μs
■ Voltage noise: ~15 nV/√Hz
But... it is designed for bipolar supply VCC = ±15VDC and shows derated performance from ±5 V. Here we have 12V max (-6/+6 virtually). It can work, it works, but not as nicely as it can. That is why I changed it for OPA1644:
■ Voltage noise: 5.1 nV/√Hz at 1 kHz
■ Gain bandwidth product: 11 Mhz (3.2 Mhz at full power)
■ Distortion: 0.00005% at 1 kHz
■ Slew rate: 20 V/μs
■ Rail-to-rail Output
■ Wide Supply Range: ±2.25 V to ±18 V
Other opams I would like to try are OP484 and OPA4227/4228. But I suppose OPA1644 will be the best one.
One more photo and apologize for probably bad English.
Some more notes and considerations:
1. Power matters. It is better to avoid using switching power supplies (PS) except the best ones that expensive as hell. Use linear power supplies. Another issue is the design of linear PS. It should have low output impedance not only at DC point (constant current), but at some frequencies range. Commonly used regulators like LM317 aren't the best solutions. However, they are better than switching PS.
2. Input caps. Though you can be sure that your sources have no bias, you can't say anything definite about your friend's sound card/dac/mobile phone/etc. I am not sure about all possible sources so caps are in their places.
3. Small film caps all over board here and there in parallel with electrolytic caps. They can both deliver power really quickly and filter noise, and form parasitic LC resonance contours. Board track longing 30 mm has the inductance roughly about 30 nH, this gives the resonant frequency about 0,9 MHz with 1 uF cap. It is far beyond audible range but some induced sub-harmonics can be audible. So, the behavior of the system should be measured, not estimated. There can be surprises. Is it really necessary to use such caps? What are pros and cons? I preferred to do without them.
4. Oxidized RCA block. It deserves changing, but I can't find a suitable good block on the board. Apparently, I'll have to put the connectors on the panel.
). God knows why the owner got rid of it for this amount of money. It was in mint stock condition.Since I'm a bit of an engineer, I decided to get into it with a soldering iron. Board routing is nice, no SMD components, it is very easy to modify and repair the board.
Stock solder, as far as understand, is trendy leadless substance. It is sh*t that is prohibited for use in hirel products in Russia. I want my amplifier to be reliable, that is why I use old-fashioned russian POS-61 (61% Sn / 39% Pb) or WBT-0800. This amp was soldered with WBT. I will not throw out it so I do not worry for the nature.
Power protection diode still in place, 2х 1000 uF caps are changed to Panasonic FR 1500 uF 25 V, FR 470 uF 25 V near them. Another FR 1500 uF 25 V near opamp. These caps are rather small, cheap, has low ESR and high endurance. That is why I use them in power circuits.
Then 6x 220 uF 16 V Nichicon FG near opamp and 2x 10 uF 50 V Nichicon FG. Probably not the best ones, but better than stock. Ceramic caps are changed to 220 pf WIMAs, 2x matched 5,1 KOhm russian MLT (metal film) gain resistors.
Input caps - Nichicon ES 4.7 uF 50 V, output caps - Nichicon ES 470 uF 35 V. Amp drives 250 Ohm DT770Pro headphones, thus 470 is enough regarding proper cutoff frequency. It would be better to use 1000 uF 16 V at the output, but I didn't find them in stock in Moscow, and my impatience said "use 470" Electrolytic caps at the input is not the best solution, but I do not want to omit them. There is not enough space to place PP foil caps, maybe I will try PS/PET coated caps later.
Output stage is ok, no mods here.
Now move on to the next part - opamp. ST MC33079 in stock. According to the datasheet:
■ Voltage noise: 4.5 nV/√Hz
■ Gain bandwidth product: 15 MHz
■ Slew rate: 7 V/μs
■ Distortion: 0.002%
Not so bad, frankly speaking. But it was changed for IC holder. I tried OPA404 (on pics), it's nice old amp:
■ Gain bandwidth product: 6.4 MHz
■ Slew rate: 35V/μs
■ Voltage noise: ~15 nV/√Hz
But... it is designed for bipolar supply VCC = ±15VDC and shows derated performance from ±5 V. Here we have 12V max (-6/+6 virtually). It can work, it works, but not as nicely as it can. That is why I changed it for OPA1644:
■ Voltage noise: 5.1 nV/√Hz at 1 kHz
■ Gain bandwidth product: 11 Mhz (3.2 Mhz at full power)
■ Distortion: 0.00005% at 1 kHz
■ Slew rate: 20 V/μs
■ Rail-to-rail Output
■ Wide Supply Range: ±2.25 V to ±18 V
Other opams I would like to try are OP484 and OPA4227/4228. But I suppose OPA1644 will be the best one.
One more photo and apologize for probably bad English.
Some more notes and considerations:
1. Power matters. It is better to avoid using switching power supplies (PS) except the best ones that expensive as hell. Use linear power supplies. Another issue is the design of linear PS. It should have low output impedance not only at DC point (constant current), but at some frequencies range. Commonly used regulators like LM317 aren't the best solutions. However, they are better than switching PS.
2. Input caps. Though you can be sure that your sources have no bias, you can't say anything definite about your friend's sound card/dac/mobile phone/etc. I am not sure about all possible sources so caps are in their places.
3. Small film caps all over board here and there in parallel with electrolytic caps. They can both deliver power really quickly and filter noise, and form parasitic LC resonance contours. Board track longing 30 mm has the inductance roughly about 30 nH, this gives the resonant frequency about 0,9 MHz with 1 uF cap. It is far beyond audible range but some induced sub-harmonics can be audible. So, the behavior of the system should be measured, not estimated. There can be surprises. Is it really necessary to use such caps? What are pros and cons? I preferred to do without them.
4. Oxidized RCA block. It deserves changing, but I can't find a suitable good block on the board. Apparently, I'll have to put the connectors on the panel.
You are much better off using 60/40 (lead / tin) solder..... easier to work with and lasts a lifetime..... the unleaded stuff "may" comply with ROHS (Green directive) but it is pure rubbish. Critical components (aircraft / miltary and medical equipment) are exempt from the ROHS directive and they still use 60/40 solder which tell you a lot about how crap unleaded solder is.
I had to solder SOIC14 on the SOIC-DIP board just as you do. Opamp on photos - OPA404KP.
