Hi Banet, good to hear from you. I often do point to point to avoid time and cost of devellopent work. I feel a mixture of PCB and dead bug is good.
For fun about a years ago I built a primitive transistor amp to show the forum. I was so surprised how well it worked ( RS tag strip ). For the first time in my life I had to keep it simple beyond even my standards. If not is would oscillate. It was a rats nest. I just worked so well. Most of my phono stages are dead bug . Dead bug is even more correct than point to point. I like MC33079 as it is a great chassis for that. Even so I have a few tricks not seen on many.
Dejan . Your city is full of fanatics.
This weeks project is a very crude regenerator to replace a variac. Old PA amp +/- 53 V 500 VA PSU. Compimentary feedback pair with monster gain. 1N4007 biased in dead bug expoxy glued to the outputs . Transfomer with variable resistor to supply drive. Output via a 0-12 0-12 connected 24V. No feedback, distortion as the mains ( circa 4% ). The advantage is I can set to 0.1V without any trouble. If I blow a set of transistors that's about all it will cost. The feedback pair is prefered as the bias is so easy to set. Re will be related to the 1N4007 needs. I can set exact bias by current flow ( 1N4007 Vak set by current). The 1N4007 is about an exact match in this application. Standing current will be 5 mA to 50 mA hot.
For fun about a years ago I built a primitive transistor amp to show the forum. I was so surprised how well it worked ( RS tag strip ). For the first time in my life I had to keep it simple beyond even my standards. If not is would oscillate. It was a rats nest. I just worked so well. Most of my phono stages are dead bug . Dead bug is even more correct than point to point. I like MC33079 as it is a great chassis for that. Even so I have a few tricks not seen on many.
Dejan . Your city is full of fanatics.
This weeks project is a very crude regenerator to replace a variac. Old PA amp +/- 53 V 500 VA PSU. Compimentary feedback pair with monster gain. 1N4007 biased in dead bug expoxy glued to the outputs . Transfomer with variable resistor to supply drive. Output via a 0-12 0-12 connected 24V. No feedback, distortion as the mains ( circa 4% ). The advantage is I can set to 0.1V without any trouble. If I blow a set of transistors that's about all it will cost. The feedback pair is prefered as the bias is so easy to set. Re will be related to the 1N4007 needs. I can set exact bias by current flow ( 1N4007 Vak set by current). The 1N4007 is about an exact match in this application. Standing current will be 5 mA to 50 mA hot.
With all harmonics (far) below -90dB, individual differences are so small, that it doesn't really matter if there's a gradual decrease of higher harmonics or not. In which case, also Mr Hiraga should/would be pleased.
The human ear can not sense pressure, only pressure differences, but not on a linear scale.
Same same for harmonics, and their distribution, have all of them small enough and the ear can not detect.
The first zero is to be under the THD threshold radar, additional ones are not a primary target, but a means to achieve the second goal for transparent sound.
Diffferent approaches to achieve the same goal, current sota parts enable one approach to reach the objective much easier and with fewer components than the alternative.
Amusing bit, in the end, is that both rely on decent bias current levels, buffers or output stages.
Last edited:
That's what I have been saying for a long time. In terms of audio fanatics per capita, I think we're up there near the top.
Strictly speaking. Banatm is from Pančevo, a sizeable town some 16 km (app. 10 miles) from Belgrade. From my window, I oversee it every day.
From my home, say 15-20 minutes of an easy drive with no speed limit breaking. In between is an oil refinery and some petrochemical facilities, but the gap is narrowing day by day. Against my best wishes, mind you, for the better of life it would be best if Pančevo remained as it is, famous for easy going people, who are reputed to be the worst drivers in all of even former Yugoslavia.On fanaticism - I must make it clear, this is not blind, deaf and dumb drive God knows where, this is I think true grit, people who really believe in what they are doing. They honestly try to do better. But this is some way above to what you would call pundits, although not really fanaticism.
Nigel
Maybe by using 3D PTP wiring layout with only solid core Cu.wires we can avoid three main non-desirable effects ,
- first effect is to avoid electromagnetic induction from OPS lines loops loaded with dynamic changing heavy currents to influence input stages with relative small currents loops , which can be in order of uA or even nA .
-second one is to avoid forming of parasitic capacitance of dense packed 2D PCB layout and in the same time to avoid capacitive parasitic influence from low impedance loops into relative high impedance loops .
And third effect which is not related only for good designed PTP layout but for good designed PCB layout to , is to avoid direct mixing of heavy DC currents inside of OPS loops with relative low DC current loops of input , VAS , and drivers stages .
After all Audio Amps are not some RF units which work in VHF/UHF/GHz region but usually work in linear region from DC to 350Khz.
Last edited:
That's what I have been saying for a long time. In terms of audio fanatics per capita, I think we're up there near the top.
Strictly speaking. Banatm is from Pančevo, a sizeable town some 16 km (app. 10 miles) from Belgrade. From my window, I oversee it every day.
On fanaticism - I must make it clear, this is not blind, deaf and dumb drive God knows where, this is I think true grit, people who really believe in what they are doing. They honestly try to do better. But this is some way above to what you would call pundits, although not really fanaticism.
Wrong
I do not live an I never lived in Pancevo but in one small village about 33Km from Pancevo , and I don`t like life in the big cities as Belgrade or Pancevo from some very practical reasons , one of them is that if I want to listen some very loud music in my not so small house just in middle of the night I can do that without to bother anybody , since residents density is very small here .
I heard about that legend that people from Pancevo are very bad car drivers , I don`t know is that truth or not , but I found that even with my almost 54 age I can still perfectly control my heavy tuned GSXR750J at 270 - 300Km/h ,or driving only on the rear wheel for kilometres at 200Km/h .
Many of the early Japanese amplifiers/receivers were wired point to point. It's amazing to see the insides of those units, so much hand work.
Yes it is just nice to see inside of all that`s early units ! , but today that hand wired PTP human work will be unacceptable expensive way of units production for HI-FI industry , only we DIY people and just some small exotic audio companies can afford that PTP approach .
Last edited:
I really do nothing it is a good idea to use the point-to-point wiring for transistors, IC's and SMD devices
Hi Pavel !,
I can understand your concerns ,
but I have one question for you :
Did you ever ask yourself what shapes stray electromagnetic fields have around of piece of straight solid core Cu wire , and on the other hand around of flat PCB Cu track full of sharp edges ,where both type of conductors conducts let`s say some heavy currents of power SS OPS devices in some SS power amp ?
Best Regards !
@Banat
Well, your avatar on the local DIY forum namens Pnačevo as your place of residence. I took it from there.
On your PTP notes, without arguing, I'd like to point out some facts.
Any designer worth half his salt will know what he faces with PC boards. They make things MUCH faster and easier and generally more pratcical, but as ever, there is a price to pay. Knowing most of the problems you have to face, you can:
1, Use thicker than usual copper layers; standard in commercial goods is 30-35 microns, I like to skip the 50 micron group and go straight to the 70 micron group. This significantly reduces stray capacitance;
2. You have to make sure you group the components as you should, avoid mixing low level with high level components (e.g. input stage with output stage);
3. Try to group all your components in a block as close together as is practical;
4. Have separate ground leads for input signal, NFB return signal (do NOT mix them together, better have them separate) and for each leg of Vcc/Vee;
5. Use "clean ground" methods, and
6. Use two sided circuit boards, making sure that high power line are as thick as possible, and dedicate whatever is left over to ground, as is usually done in RF technology.
I realize this is considerably more work in planning the layout, but it never fails to pay dividends. Also, this will reduce your RF succetability to level lower than you get with PTP, unless you take the time and trouble to isolate each and every leg of whatever you want to connect. A lot of bare wire is an invitation to trouble, always.
Well, your avatar on the local DIY forum namens Pnačevo as your place of residence. I took it from there.
On your PTP notes, without arguing, I'd like to point out some facts.
Any designer worth half his salt will know what he faces with PC boards. They make things MUCH faster and easier and generally more pratcical, but as ever, there is a price to pay. Knowing most of the problems you have to face, you can:
1, Use thicker than usual copper layers; standard in commercial goods is 30-35 microns, I like to skip the 50 micron group and go straight to the 70 micron group. This significantly reduces stray capacitance;
2. You have to make sure you group the components as you should, avoid mixing low level with high level components (e.g. input stage with output stage);
3. Try to group all your components in a block as close together as is practical;
4. Have separate ground leads for input signal, NFB return signal (do NOT mix them together, better have them separate) and for each leg of Vcc/Vee;
5. Use "clean ground" methods, and
6. Use two sided circuit boards, making sure that high power line are as thick as possible, and dedicate whatever is left over to ground, as is usually done in RF technology.
I realize this is considerably more work in planning the layout, but it never fails to pay dividends. Also, this will reduce your RF succetability to level lower than you get with PTP, unless you take the time and trouble to isolate each and every leg of whatever you want to connect. A lot of bare wire is an invitation to trouble, always.
@Banat
Let me tell you, the legend is not a legend, it's the truth. In Belgarde, we see a lot of PA registration plates.
A few days ago, a guy in a brand new FIAT with Pančevo plates, driving in a traffic jam (intersection of Takovska and Lole Ribara), opened his window and threw out a cut wide open matt black garbage bag. It stuck straight across my windscreen and I was completely blind in less than a second. I jammed the brakes, and a good thing I did because he did as well, but as I struggled to take the damn vinyl off my front window, the cars started moving. As usual, no cop to be seen anywhere.
Now, you tell me, how would you call that kind of driving?
Let me tell you, the legend is not a legend, it's the truth. In Belgarde, we see a lot of PA registration plates.
A few days ago, a guy in a brand new FIAT with Pančevo plates, driving in a traffic jam (intersection of Takovska and Lole Ribara), opened his window and threw out a cut wide open matt black garbage bag. It stuck straight across my windscreen and I was completely blind in less than a second. I jammed the brakes, and a good thing I did because he did as well, but as I struggled to take the damn vinyl off my front window, the cars started moving. As usual, no cop to be seen anywhere.
Now, you tell me, how would you call that kind of driving?
The path to convincing sound can be via many routes - a speaker that is engineered precisely to suit the typical amplifier is able to make it happen ... the comments by the user of these studio monitors, https://www.gearslutz.com/board/high-end/949764-amphion-beautiful.html, are ticking all the right boxes ...
Not quite as bad as that. These are not malicious people, or people with vendetta on their minds, much less road rage. These are usually very easy going people, far from aggressive, but capable of doing silly things because they are too slow rather than being too fast.
True, but in my view, no part of the audio chain should be picky about what feeds it or drives it.
I know that's easier said than done, but it can be done if one is peristent enough.
I find that my present system delivers a consistent sound and the only differences are brought about when I change the speakers. Best results are achieved with biamping, but unfortunately the AR94 speakers cannot be biamped, just one pair of binding posts terminals.
Point to point wiring is a nice reality if a power supply to power amp layout. My first experiment was to take an off the shelf power amp module and use the spare wire stump of the TO3 to as my terminal. Instead of the power going in via the PCB it went to the device. Simple and the best it can be. Even the other devices seem better served by this. In my ideal world a small RC filter to the VAS and loose a bit of book spec on maximum power. 220 R and 2200 uF comes to mind. Realistically 3 V would be the useful loss. If using non PCB PSU on the analyzer one then bends the wires for best noise and other effects, 1 cm and 10 degrees can be a very big deal. Rotate the toroid also as the magnetic field is usually greater where the wires come out. I often use 2.5 mm house wiring cable as it is cheap and works well for this. As said previously the decoupling should deal with quallity of the PSU and inductance problems. Most amps I see have the same thinking as custom motorcars. It must look good. Problem is this usually corrupts the design. Keep It Simple Stupid ( KISS ). Same with cars. Big tyres + water = death. Simple reason is the tyre is wrong. A tyre can have optimum wet grip. Not too small and not too big. In the dry usually it will have more than enough grip if the wet grip problem is adressed. The reason is simple . The tyre works at different speed as the wheel turns inside to outside. In the end the tyre hysterisis can not compensate. If the wheel was two tyres with a differential it would be better. Power amplifers have an optimum capacitance. Valve designers are forced to know this as they start to run out of choices. They even notice small is beautiful sounding. The advantage valve people have is ripple is a constant. It is easy to know if the amp works. Transistor amps hide the ripple. No one discovers this as they are not thinking about it. Simple answer is to forget the sceince of the transistor amps and pretend it is a valve amp. The VAS and LTP must be fed with DC and not DC + ripple. When so the LTP can use a simple resistor. My analogy is we are supplying water to the house and taking away the toilet waste water in the same pipe. We then filter the drinking side. Crazy. Even more crazy is we don't even filter it!
Most transistor amps have dreadfull power supplies. Most transistor amplifiers assume this and have various design features to overcome ripple. This offers the oppertunity to do point to point PSU's. That is 10 000, 22 000, 47 000 uF caps and a big rectifier. The quality not really an issue as the decoupling can assist ( 2200 uF per output device , high grade). To me that's point to point for all. The alternative is more difficult and costs money.
AR 94 bi-amp. Should be easy. One way is build an active tweeter amp. Personaly I would use capacitor coupling as the device will be filter No1 and protection in one ( you might need more poles ). A high grade 10 uF can be 100 V polyester or whatever is better. Build the amp with a bootstraped input to make it have a very high input Z. If classA the bootstrap should be benign . It goes between op amp minus and 2 x 10K on the input. Usually a series resistor and capacitor ( 5 K at 1 kHz - 3 dB ? ). The the little amp can be in paralell with the preamp output. I am assuming an OPA2604 with booster transitors. Myself I would use complimentary feedback pairs with 200 mA standing current. I would not have the outputs in the feedback loop as the tweeter will not upset the working and will stay in class A well enough. After listening tests feedback can be tried. The op amp will have traditional feedback, be it direct or via the output stage. Can have 50/50. OPA 2604 was very stable used this way.
Most transistor amps have dreadfull power supplies. Most transistor amplifiers assume this and have various design features to overcome ripple. This offers the oppertunity to do point to point PSU's. That is 10 000, 22 000, 47 000 uF caps and a big rectifier. The quality not really an issue as the decoupling can assist ( 2200 uF per output device , high grade). To me that's point to point for all. The alternative is more difficult and costs money.
AR 94 bi-amp. Should be easy. One way is build an active tweeter amp. Personaly I would use capacitor coupling as the device will be filter No1 and protection in one ( you might need more poles ). A high grade 10 uF can be 100 V polyester or whatever is better. Build the amp with a bootstraped input to make it have a very high input Z. If classA the bootstrap should be benign . It goes between op amp minus and 2 x 10K on the input. Usually a series resistor and capacitor ( 5 K at 1 kHz - 3 dB ? ). The the little amp can be in paralell with the preamp output. I am assuming an OPA2604 with booster transitors. Myself I would use complimentary feedback pairs with 200 mA standing current. I would not have the outputs in the feedback loop as the tweeter will not upset the working and will stay in class A well enough. After listening tests feedback can be tried. The op amp will have traditional feedback, be it direct or via the output stage. Can have 50/50. OPA 2604 was very stable used this way.
At college electronics was taught by Peter Platt. He retired from the RAF to do teaching. Peter was hopeless on things we who loved electronics already knew. He was great on the difficult stuff. I was told off repeatedly for asking questions. That was until other RAF guys insisted he answered them. They knew the stuff and were there to get a piece of paper. One day I asked about Kirkhoff. Peter said something stunning ( 1972 ). " That's the problem with you audio guys you speak of the signal path. No such thing. It is all in the signal path". When a problem gets too big one solves it by using Gaussian principles. That is we call an amplifier a node and only ask ourselves about roads in and roads out ( wires ). What happens in the city is not in our control. This would be an op amp or chip amp. Also a tightly built PCB amp. Simple design allows us to turn the problem back into Kirkhoff. Sure CAD can say how to do it. Problem is CAD will ask you to do it the way it knows. Same with cars. This length , this Cd = this car.
This is why asking what is the lowest distortion we need is important. Any fool can see an amplifier that goes all out for low distortion will almost certainly be A Gauss rather than Kirkhoff problem.
BTW. Class A SE valve amp distortion and class AB/D distortion are not the same. One can not infer anything from setting a bench mark. 1% THD valve SE and 0.01% class AB might be correct for both. 0.1% transistor class A also. With class AB look very hard and you almost certainly will find bad things regardless of the low THD. In fact relaxing the feedback will often get rid of probblems. There will be an optimum which might seem poor on paper. Op amps certainly are like this. My rule is Gain 1 often is good. > gain 10 is good. Anything else be careful.
Spell checker is not working so accept my version of these words.
This is why asking what is the lowest distortion we need is important. Any fool can see an amplifier that goes all out for low distortion will almost certainly be A Gauss rather than Kirkhoff problem.
BTW. Class A SE valve amp distortion and class AB/D distortion are not the same. One can not infer anything from setting a bench mark. 1% THD valve SE and 0.01% class AB might be correct for both. 0.1% transistor class A also. With class AB look very hard and you almost certainly will find bad things regardless of the low THD. In fact relaxing the feedback will often get rid of probblems. There will be an optimum which might seem poor on paper. Op amps certainly are like this. My rule is Gain 1 often is good. > gain 10 is good. Anything else be careful.
Spell checker is not working so accept my version of these words.
Well, SMD devices would be a Herculean challenge to solder point to point.
1206's are not too bad , nor BCV61/62. I use the round ones from what was Philips. I have my Parkinsons, so if I can anyone can. How I do it is like the sea going in and out. I wait for the moment.
Last night I did some plumbing, I knew if I didn't it might be 6 months to get up the courage. It was exactly the trials of Mr H. The pipes sounded like when something is wrong on the car due to the hands and spanners. I then made the big leap forward. I used my left hand. It was better. I started at 9 PM and had many problems. Finished at what I thought was 1 AM. It was 11 05 PM . The dam gate valves are finally useless so it was a complete drain down. Went to bed very happy. If I didn't know it impossible I would have said time stood still as it seemed longer. My doctor thinks my version will not change too much with time. If so I will be very happy. 3 years ago I thought my world would have to change. Gradually I have found ways to do things. I have days when I can write with a pen, mostly not. These days who cares? When the pipes made the rattle noise I just laughed. First time I ever could. The best thing was to work out how best to get to the job. I was working in a tight space and didn't drop a thing. That would be unusual for anyone.
- Status
- Not open for further replies.