Production in China or India will be cheaper – but probably you'll have to be there physically, to make sure it's being done according to specs …
If it were only so easy, Joshua. Do you have any feeling for how many chassis per year that I am indirectly responsible for that are made in China and Taiwan?
Of course, American made costs more, because Americans have bigger personal expenses in order to support themselves and their families. If you don't make MANY of the same chassis, the Chinese just won't do it. Why should they?
Of course, American made costs more, because Americans have bigger personal expenses in order to support themselves and their families. If you don't make MANY of the same chassis, the Chinese just won't do it. Why should they?
If I thought you were seriously considering lower cost high quality chassis manufacture I'd point you in the right direction. But since it's just academic...
OK, back to noise and noise cancellation/attenuation. It seems to me that there are two schools of thought. a) Do noise attenuation in the power supply, and b) do some of it in the psu, and some of it in the circuit itself. Examples of the latter are using CCSs as loads, opting for push-pull topologies, or something like the aikido circuits.
I'm probably leaving some out, but you the experts can fill in and correct me.
Option b) implies a circuit with good psrr. syn08 has repeatedly said that this is good engineering, and that option a) is bad engineering. I don't want to single syn08 out, but he's been the most recent to say this and most vocal about it.
One of the advantages of being a beginner, like me, is that I start from zero. Even better, I'm not an engineer.
(you can start shooting now)
So naively I think: hm... I like things to be separate. I read all sorts of threads about amplifier design and get the sense that it's very difficult to keep things clean and simple. Especially to keep a signal undistorted. And then I see that, for instance, the aikido circuit has a noise canceling stage for each amplification stage. Right in the middle of amplifying things, there it is, the thing trying to manipulate the signal to take out only the noise, and leave the signal intact.
How about the single-ended with CCS load, that gives that so much wanted psrr? It does look a little like part of the regulator/filter moved very close to the amplifiers (and it also became the load, while at it). Other people take it out and put it in the psu, add a capacitor or two to it, and call it a capacitor multiplier, or a gyrator, or whatever.
I wont' even touch the push-pull thingy. It's something that I don't feel good about.
Is there more to the picture? Am I missing something really obvious here? Dear experts, please be kind to a beginner.
OK, back to noise and noise cancellation/attenuation. It seems to me that there are two schools of thought. a) Do noise attenuation in the power supply, and b) do some of it in the psu, and some of it in the circuit itself. Examples of the latter are using CCSs as loads, opting for push-pull topologies, or something like the aikido circuits.
I'm probably leaving some out, but you the experts can fill in and correct me.
Option b) implies a circuit with good psrr. syn08 has repeatedly said that this is good engineering, and that option a) is bad engineering. I don't want to single syn08 out, but he's been the most recent to say this and most vocal about it.
One of the advantages of being a beginner, like me, is that I start from zero. Even better, I'm not an engineer.
(you can start shooting now)So naively I think: hm... I like things to be separate. I read all sorts of threads about amplifier design and get the sense that it's very difficult to keep things clean and simple. Especially to keep a signal undistorted. And then I see that, for instance, the aikido circuit has a noise canceling stage for each amplification stage. Right in the middle of amplifying things, there it is, the thing trying to manipulate the signal to take out only the noise, and leave the signal intact.
How about the single-ended with CCS load, that gives that so much wanted psrr? It does look a little like part of the regulator/filter moved very close to the amplifiers (and it also became the load, while at it). Other people take it out and put it in the psu, add a capacitor or two to it, and call it a capacitor multiplier, or a gyrator, or whatever.
I wont' even touch the push-pull thingy. It's something that I don't feel good about.
Is there more to the picture? Am I missing something really obvious here? Dear experts, please be kind to a beginner.
I know, John, I was joking. In me previous job I've dealt with Chinese mold makers and I know it isn't worth any ones while for small quantities.
ikoflexer, not being an engineer myself, but having a lot of experience as an electronics technician and an audiophile, these are my simplistic thoughts on the matter:
First, any amplifier would better be designed with highest sound quality as the major goal. PSR would better be considered, however, only as a second goal.
Second. Once the amplifier design is complete, the needs from the PSU should be looked at and the PSU should be designed accordingly.
Generally speaking, the quieter the PSU is, the better – as long as line regulation and load regulation are maintained at appropriate level and as long as its' production cost is reasonable, or within the overall project's cost. Once the amplifier's PSR is known, the degree of possible compromise in the PSU's design is known as well.
So, it may be an overkill to design a super quiet (and expansive) PSU without knowing the amplifier's requirements. Also, there are different considerations for a class A amplifier, vs. class AB one.
First, any amplifier would better be designed with highest sound quality as the major goal. PSR would better be considered, however, only as a second goal.
Second. Once the amplifier design is complete, the needs from the PSU should be looked at and the PSU should be designed accordingly.
Generally speaking, the quieter the PSU is, the better – as long as line regulation and load regulation are maintained at appropriate level and as long as its' production cost is reasonable, or within the overall project's cost. Once the amplifier's PSR is known, the degree of possible compromise in the PSU's design is known as well.
So, it may be an overkill to design a super quiet (and expansive) PSU without knowing the amplifier's requirements. Also, there are different considerations for a class A amplifier, vs. class AB one.
Everyone contributing to chassis construction. I appreciate each and every input that gives some information on how to do it better or cheaper, or both. However, off hand remarks like: I could get it done for 1/3 the price, doesn't float my boat, or actually give anyone else any meaningful info, either. Joking is not part of the package.
John
what happened to your welding the Aluminium chassis try out. Welding would have avoided using large mass for the milling out substantial amount of material. Ofcourse the welding method is specific to Alumunium. It is not easy like welding steel - requires TIG /Plasma or Laser welding under inert atmosphere like Argon- expensive. This process would have saved cost in my opinion or your experience is otherwise?
kannan
We tried welding with the first prototypes. In fact, Jack Bybee has one of these early cases, surrounding his CTC preamp. However, it was too difficult to do it right, so we went to 'hogging' the cases out.
We have made many cases over the years in the USA, where my associates and I were responsible for making them properly. We tried plastic, sheet metal, machined sides, aluminum welding, and finally hogging out. Each has its good points, and bad points. However, nothing comes for free, and really good does not come cheap.
We have made many cases over the years in the USA, where my associates and I were responsible for making them properly. We tried plastic, sheet metal, machined sides, aluminum welding, and finally hogging out. Each has its good points, and bad points. However, nothing comes for free, and really good does not come cheap.
I might say something specific and important about the price and quality of preamplifier and amplifier cases. I am NOT a machinist, or a mechanical engineer, but I have made audio products that require custom cases for the last 35 years. In fact, the world has gone full circle, and the guy who designed the enclosure for the original Symmetry Crossover, 32 years ago, invited me to lunch, this week.
Now, in between, what have I learned? Well, talk is cheap, and quality control can make or break you.
The 'low bid' can well be the 'high bid' once the headaches start.
For example, when we started Vendetta Research over 25 years ago, I had hoped to use a pretty, pre-finished plastic box for the SCP1 pre-preamp. I used an internal conductive spray to create the electrostatic shield necessary. Talk about minimum thickness! Still, I had hoped to keep the prices down, and the rework to make the boxes useful, at a minimum.
Well, we outgrew the plastic box, when it was discovered by open listening test that we could hear the Tantalum electrolytic caps in the power supply bypass. Of course, some here would insist that I am imagining hearing any differences, yet some here also castigate me for using a circuit with such poor power supply rejection that the cap is essentially in the audio circuit, because that is why it was so easy to hear.
Well, it came to making a decent custom enclosure for the Vendetta Research phono stage.
One company, strongly recommended to us, was located in California, and they did audio enclosures for other audio manufacturers, already. Initially, we made some initial prototypes to try out the new package, and the initial prototypes came out almost perfectly. Then, we submitted the prototype chassis and paperwork to the highly recommended independent company to make in quantities of 50 or so, for each batch.
Now this is what happened: Sometimes the top and bottom plates were scratched, as if they had been handled in a 'dirty' environment. One time they forgot to BEND the sheet metal power supply box that makes it an enclosure, but they still finished and lettered the batch anyway, and sent them along to us.
Then, we started to have FIT problems. We had to hand match different lids to different cases, and you could see that drill holes were often offset from one side to the other. This was OUR fault, I was told, because the box that we designed, required a near perfect fit, i.e. no slop, to make it fit together. Just because we had made dozens of the same boxes previously that fit together easily and perfectly, was no reason that future cases needed to do so.
Later, with even other fabricators, I hired a mechanical engineer to make precise blueprints for submission, so at least I had a reference to argue over. Now 17 years, after Vendetta Research has ceased to need boxes, who do I meet at the RMAF, Oct 3? The son of the guy who was in charge of the company that we first used, where we first had problems, 25 years before. He is in charge now, and assures me that these sort of problems are not likely today. In fact, he still remembers the boxes that his father's company made for us. Once again. full circle. Do I trust him? Yes, but 'Trust and verify' in future ;-)
Now, can people begin to understand why one might stick to one, VERY GOOD vendor, even if they seem to have a slightly higher price, than the lowest bid, and how we might get suspicious of a VERY LOW bid?
I say all this, because 'loose talk' about subjects like this, cause confusion in others, who have not yet the experience with specifying cases, etc. First, they might think that I was intentionally ripped-off by our vendor, and that we stupidly went along, or that cheap case can be easily found without large volume production, and in-house expertise, and oversight, as to how to get the job done. Please keep these lessions in mind, everyone.
Now, in between, what have I learned? Well, talk is cheap, and quality control can make or break you.
The 'low bid' can well be the 'high bid' once the headaches start.
For example, when we started Vendetta Research over 25 years ago, I had hoped to use a pretty, pre-finished plastic box for the SCP1 pre-preamp. I used an internal conductive spray to create the electrostatic shield necessary. Talk about minimum thickness! Still, I had hoped to keep the prices down, and the rework to make the boxes useful, at a minimum.
Well, we outgrew the plastic box, when it was discovered by open listening test that we could hear the Tantalum electrolytic caps in the power supply bypass. Of course, some here would insist that I am imagining hearing any differences, yet some here also castigate me for using a circuit with such poor power supply rejection that the cap is essentially in the audio circuit, because that is why it was so easy to hear.
Well, it came to making a decent custom enclosure for the Vendetta Research phono stage.
One company, strongly recommended to us, was located in California, and they did audio enclosures for other audio manufacturers, already. Initially, we made some initial prototypes to try out the new package, and the initial prototypes came out almost perfectly. Then, we submitted the prototype chassis and paperwork to the highly recommended independent company to make in quantities of 50 or so, for each batch.
Now this is what happened: Sometimes the top and bottom plates were scratched, as if they had been handled in a 'dirty' environment. One time they forgot to BEND the sheet metal power supply box that makes it an enclosure, but they still finished and lettered the batch anyway, and sent them along to us.
Then, we started to have FIT problems. We had to hand match different lids to different cases, and you could see that drill holes were often offset from one side to the other. This was OUR fault, I was told, because the box that we designed, required a near perfect fit, i.e. no slop, to make it fit together. Just because we had made dozens of the same boxes previously that fit together easily and perfectly, was no reason that future cases needed to do so.
Later, with even other fabricators, I hired a mechanical engineer to make precise blueprints for submission, so at least I had a reference to argue over. Now 17 years, after Vendetta Research has ceased to need boxes, who do I meet at the RMAF, Oct 3? The son of the guy who was in charge of the company that we first used, where we first had problems, 25 years before. He is in charge now, and assures me that these sort of problems are not likely today. In fact, he still remembers the boxes that his father's company made for us. Once again. full circle. Do I trust him? Yes, but 'Trust and verify' in future ;-)
Now, can people begin to understand why one might stick to one, VERY GOOD vendor, even if they seem to have a slightly higher price, than the lowest bid, and how we might get suspicious of a VERY LOW bid?
I say all this, because 'loose talk' about subjects like this, cause confusion in others, who have not yet the experience with specifying cases, etc. First, they might think that I was intentionally ripped-off by our vendor, and that we stupidly went along, or that cheap case can be easily found without large volume production, and in-house expertise, and oversight, as to how to get the job done. Please keep these lessions in mind, everyone.
Last edited:
This is in accord with my own experience as purchasing manager for electronics manufacturing company.
John,
Making sheet metal oncie's is easy, each piece is fit to match. In large quantities you can tune in the equipment to make all the pieces the same.
Small quantities are a B... When you bend sheet metal it gets longer by about 40% of the material thickness. It depends on the tensile strength and the absolute precise setting of the bending pressure or distance limits. The tensile strength in a roll of steel is not very uniform in the first few or last few feet, so you have to only use the middle. Aluminum is more often obtained in sheets so you do not know which part of the roll it came from. Changes in the mill making it have a big difference in strength. Then the thickness also varies. This is a double whammy.
As if the material difference is not enough, lets talk about tooling. I can buy precision forming tools or "standard." The precision is interchangeable. So if you order 50 today and 50 after my tooling has been changed.... Then my press brake is level within .005" at 8' in the morning, in the afternoon sun it is different by .015"
I have however seen absolutely beautiful sheet metal products where everything not only fits, but snap fits like the bud boxes do, only without bud's gaps.
Sheet metal work has been around a long time, computer controls and optical sensors make it better. Skill still is best.
As to audio cases, I have built heatsinks out of laminations, 198 of them to be exact. Gives me lots of radiating area and good control of fin to gap ratios. Cases can be built the same way, think of log homes. Of course that method was a military secret in WWII. It is how we built radar tubes at low cost in great quantity.
Making sheet metal oncie's is easy, each piece is fit to match. In large quantities you can tune in the equipment to make all the pieces the same.
Small quantities are a B... When you bend sheet metal it gets longer by about 40% of the material thickness. It depends on the tensile strength and the absolute precise setting of the bending pressure or distance limits. The tensile strength in a roll of steel is not very uniform in the first few or last few feet, so you have to only use the middle. Aluminum is more often obtained in sheets so you do not know which part of the roll it came from. Changes in the mill making it have a big difference in strength. Then the thickness also varies. This is a double whammy.
As if the material difference is not enough, lets talk about tooling. I can buy precision forming tools or "standard." The precision is interchangeable. So if you order 50 today and 50 after my tooling has been changed.... Then my press brake is level within .005" at 8' in the morning, in the afternoon sun it is different by .015"
I have however seen absolutely beautiful sheet metal products where everything not only fits, but snap fits like the bud boxes do, only without bud's gaps.
Sheet metal work has been around a long time, computer controls and optical sensors make it better. Skill still is best.
As to audio cases, I have built heatsinks out of laminations, 198 of them to be exact. Gives me lots of radiating area and good control of fin to gap ratios. Cases can be built the same way, think of log homes. Of course that method was a military secret in WWII. It is how we built radar tubes at low cost in great quantity.
I remember when Norsk Hydro introduced their satellite dishes. Afair they had 90cm diameter and where milled out of solid aluminium blocks.
I haven't read any complaints then and I find all the moaning about prices in the blowtorch thread astonishing immature.
The blowtorch preamp is a luxury product and it costs what it costs.
Regards
I haven't read any complaints then and I find all the moaning about prices in the blowtorch thread astonishing immature.
The blowtorch preamp is a luxury product and it costs what it costs.
Regards
John,
The last product of your design I bought was an ML JC2 preamp. It was sheet metal as I recall. I read your post and from the hole mistakes and mention of unbent cases thought you were into more commodity sheet metal parts.
If you have an NC milling facility that cannot match pieces that is entirely a different issue. There is some thermal dimension changes from high speed milling. But not keeping holes to a reference is just bad work. Not taking back defective product and issuing a refund is quite telling.
With NC milling the first piece costs a bit more than the rest for programing time. There may be a setup charge for small runs, but otherwise the unit piece cost should be consistent.
Oh yeah my JC2 lasted about six months, went back for warranty repair, still not back yet.
It has been many years, I don't know if the stereo store had credit problems or if it got resold. That was 3 preamps ago. Thanks anyway.
- Status
- Not open for further replies.