As with most commercial products, in my opinion a worst case calculation and Failure Mode Effect Analysis (FMEA) should be peformed ensuring the design's integrity for extended use under all known operating circumstances and environments.
MTBF (Mean Time Between Failure) analysis is useful in predicting the product's theoretical reliability.
Most half decent simulators would perform these functions automatically as well as EMC analysis of the final PCB layout. I have not seen a single question regarding electromagnetic compatibility except for KeanToken mentioning that he has no electrical earth, but no one even regarded this as an issue.
I fully realise that this is not a commercial design but it would be useful to Hugh should he want to turn this into a commercial product. It would also be very useful for those of you who is looking into establishing yourself as a future entrepeneur designing and manufacturing commercial products.
Try it guys it may be an interesting excercise.
Nico
MTBF (Mean Time Between Failure) analysis is useful in predicting the product's theoretical reliability.
Most half decent simulators would perform these functions automatically as well as EMC analysis of the final PCB layout. I have not seen a single question regarding electromagnetic compatibility except for KeanToken mentioning that he has no electrical earth, but no one even regarded this as an issue.
I fully realise that this is not a commercial design but it would be useful to Hugh should he want to turn this into a commercial product. It would also be very useful for those of you who is looking into establishing yourself as a future entrepeneur designing and manufacturing commercial products.
Try it guys it may be an interesting excercise.
Nico
This is most interesting, Nico. Only the full featured commercial simulators have this function, do you have advice on what to use?
Tomorrow I shall have tone control and power supply schematics and with amp schemat under consideration now the group should be up tomorrow. 10pm here......
Cheers,
Hugh
Tomorrow I shall have tone control and power supply schematics and with amp schemat under consideration now the group should be up tomorrow. 10pm here......
Cheers,
Hugh
High All,
I think most commercial Schematic Capture/simulation and PCB design packages have add-ins to perform almost every time consuming repetitive engineering function. However, one cannot remove the engineers abillity to define, acquired skill and logic out of the equation.
I am familiar with two products that have these capabilities, CADStar and Easy PC, there are other as these products are very competitive all competing for some market share.
Worst case calculation involves calculating every component power requirements under worst case conditions and then de-rating it accordingly to the manufacturers recommendation. I have noticed that mostly output stages carry some importance at DIY.
FMEA is a logical and formatted procedure of questions and answers, what if, then what is the failure mode and what effect will it have on the consumer, then how can it be avoided.
KT has suggested one such failure mechanism, what if the the power supply charges non-symmetrically, will it create a pop in the headphones, will the headphones be damaged, will the listeners ears be damaged, will the damaged headphones damage the amp, how can we prevent this from happening if any of these possibilities exist.
This is just an example but in control circuitry, we always make sure that something fails safely if it does fail. For instance if a transistor drives a relay and the driving circuit fails, the relay must be in a predictable state when such failure occurs, so we use a pull-down resistor on the driver transistor. We often use resistors as fuses in power circuits and I have seen Hugh do this.
EMC is harder to predict but if one follows known and logical rules than the end result is often close enough to what legislation requires to only add a components or move at track in order to comply.
Most European countries will not allow the distribution of product into their environments without it complying to regulations. I have seen toroidal transformers radiating such a high magnetic field that it interferes with a CRT display. Some amplifiers are A.M. receivers or start oscillating wildly and burn out speakers when a cellular phone is operated near them.
I am honestly not saying we should go to this extreme with our hobbies but it is worth the thought if your intension is commercialising your product. Complying with your own countries requirement is already an issue, complying with a world standard is another issue all together.
This was just a thought guys not aimed at anyone or anything in particular, but a useful exercises in engineering of a product.
Nico
I think most commercial Schematic Capture/simulation and PCB design packages have add-ins to perform almost every time consuming repetitive engineering function. However, one cannot remove the engineers abillity to define, acquired skill and logic out of the equation.
I am familiar with two products that have these capabilities, CADStar and Easy PC, there are other as these products are very competitive all competing for some market share.
Worst case calculation involves calculating every component power requirements under worst case conditions and then de-rating it accordingly to the manufacturers recommendation. I have noticed that mostly output stages carry some importance at DIY.
FMEA is a logical and formatted procedure of questions and answers, what if, then what is the failure mode and what effect will it have on the consumer, then how can it be avoided.
KT has suggested one such failure mechanism, what if the the power supply charges non-symmetrically, will it create a pop in the headphones, will the headphones be damaged, will the listeners ears be damaged, will the damaged headphones damage the amp, how can we prevent this from happening if any of these possibilities exist.
This is just an example but in control circuitry, we always make sure that something fails safely if it does fail. For instance if a transistor drives a relay and the driving circuit fails, the relay must be in a predictable state when such failure occurs, so we use a pull-down resistor on the driver transistor. We often use resistors as fuses in power circuits and I have seen Hugh do this.
EMC is harder to predict but if one follows known and logical rules than the end result is often close enough to what legislation requires to only add a components or move at track in order to comply.
Most European countries will not allow the distribution of product into their environments without it complying to regulations. I have seen toroidal transformers radiating such a high magnetic field that it interferes with a CRT display. Some amplifiers are A.M. receivers or start oscillating wildly and burn out speakers when a cellular phone is operated near them.
I am honestly not saying we should go to this extreme with our hobbies but it is worth the thought if your intension is commercialising your product. Complying with your own countries requirement is already an issue, complying with a world standard is another issue all together.
This was just a thought guys not aimed at anyone or anything in particular, but a useful exercises in engineering of a product.
Nico
Last edited:
Member
Joined 2009
Paid Member
Those are some good question Nico. Worth pointing out to the DIY people how their lives are perhaps 'too easy' !
I've always thought Double Insulating for safety is not usually recommended for DIY since any kind of fault is then unprotected. An earthed chasis is important and since Hugh has chosen a die-cast box we can ensure it's properly safety-grounded if AC line is being brought in.
I'm a bit confused this morning, must have woken up too early - are we relying on an external supply or are we still thinking of having a transformer and rectifier in the box ? - if we are using only filtering/regulation with an external supply to be provided by the constructor (e.g. wall wart) then safety is looking better.
EMC is a strong function of the box design, connectors and cables and openings in the box. I don't see that we have reason to worry about the creation of emissions unless we have a poor design (creating oscillations as you pointed out) but we could be susceptible to RF interference if somebody was living near a strong source. If I remember, Rod Elliot has some stuff on that.
Parasitic analysis of the pcb - this is pretty advanced stuff. Usually you have to pay a bit for this and know how to use and interpret the results. It's a good learning opportunity but personally I think there's enough knowledge regarding pcb layout between Hugh and Nico that this isn't worth doing. It would be nice perhaps to explain to people how/why the final layout looks the way that it does. Things like power routing, return currents, magnetic loops, separating high and low currents, trace widths and separations, size of pads and use of ground planes etc. Manufacturability is another item to consider with pcbs and there are 'rule books' on this kind of thing, although mostly for SMT. There's enough to learn that it's probably better to go pay a consultant to check the design for you - the software that can do a thorough job of this including checking that the parts selected are not near end of life, can be very expensive indeed.
The other part of manufacturability is the mechanical design and assembly. Can all the parts be soldered, are they accessible for test and repair ?
MTBF isn't so hard, we could do that by going through each part as you suggest. That's not such a bad idea even for DIY, especially for a power amp. I suspect the two most important items above all else will be voltage ratings (capacitors and semi's) along with what temperature we run things at. Capacitors are usually a dominant influence on the total system MTBF.
I've always thought Double Insulating for safety is not usually recommended for DIY since any kind of fault is then unprotected. An earthed chasis is important and since Hugh has chosen a die-cast box we can ensure it's properly safety-grounded if AC line is being brought in.
I'm a bit confused this morning, must have woken up too early - are we relying on an external supply or are we still thinking of having a transformer and rectifier in the box ? - if we are using only filtering/regulation with an external supply to be provided by the constructor (e.g. wall wart) then safety is looking better.
EMC is a strong function of the box design, connectors and cables and openings in the box. I don't see that we have reason to worry about the creation of emissions unless we have a poor design (creating oscillations as you pointed out) but we could be susceptible to RF interference if somebody was living near a strong source. If I remember, Rod Elliot has some stuff on that.
Parasitic analysis of the pcb - this is pretty advanced stuff. Usually you have to pay a bit for this and know how to use and interpret the results. It's a good learning opportunity but personally I think there's enough knowledge regarding pcb layout between Hugh and Nico that this isn't worth doing. It would be nice perhaps to explain to people how/why the final layout looks the way that it does. Things like power routing, return currents, magnetic loops, separating high and low currents, trace widths and separations, size of pads and use of ground planes etc. Manufacturability is another item to consider with pcbs and there are 'rule books' on this kind of thing, although mostly for SMT. There's enough to learn that it's probably better to go pay a consultant to check the design for you - the software that can do a thorough job of this including checking that the parts selected are not near end of life, can be very expensive indeed.
The other part of manufacturability is the mechanical design and assembly. Can all the parts be soldered, are they accessible for test and repair ?
MTBF isn't so hard, we could do that by going through each part as you suggest. That's not such a bad idea even for DIY, especially for a power amp. I suspect the two most important items above all else will be voltage ratings (capacitors and semi's) along with what temperature we run things at. Capacitors are usually a dominant influence on the total system MTBF.
Last edited:
Guy's do not get carried away, this is DIY forum and I suggested something for those dreaming or fantasising about the day that they enter the commercial area.
I think most of us do from time to time especially when we designed that great sounding amp and are passionate or even compelled to share it with the world.
This is a good place to start, while we are waiting for the final schematic to surface.
Nothing wrong keeping the thread alive with some discussion.
Gareth, you will probably be tasked to put together the BOM, besides capacitors and transistors, what about resistors? Obviously one can over specify anything, but it needs to remain commercially cost effective.
BTW you are right about deciding on a wall wart, electrically it should be safe for KT to use without being electrocuted.
I think most of us do from time to time especially when we designed that great sounding amp and are passionate or even compelled to share it with the world.
This is a good place to start, while we are waiting for the final schematic to surface.
Nothing wrong keeping the thread alive with some discussion.
Gareth, you will probably be tasked to put together the BOM, besides capacitors and transistors, what about resistors? Obviously one can over specify anything, but it needs to remain commercially cost effective.
BTW you are right about deciding on a wall wart, electrically it should be safe for KT to use without being electrocuted.
The house seems to be grounded properly from the outside, with a thick wire coming from the electrical box and attached to a rod in the ground. Perhaps it's just a decoy placed by the landlord to ward away inspectors... Maybe I should check if all the socket grounds are actually connected to each other or if they are all unconnected.
I was referring to Bigun's TGM series.
- keantoken
I was referring to Bigun's TGM series.
- keantoken
Member
Joined 2009
Paid Member
The rod is probably doing it's job. The thing to check is that it goes deep and the wire is solidly connected to it and to the earth point on your fuse panel. If you have gas pipes in the house, they should also be solidly connected to an earth point.
Actually it's the phone box that's clamped on. The spiral tube comes down from the power meter box, and a wire comes out of it attaching to the ground rod. It looks like everything's okay, but the connection to the ground rod is underground so I can't tell if it's really connected.
I know the sockets aren't grounded because I have a socket checker specifically for this purpose.
- keantoken
I know the sockets aren't grounded because I have a socket checker specifically for this purpose.
- keantoken
...
Yes, a thick silver colored wire comes through the conduit and is what attaches to the ground post. You'll have to forgive me, I've never been through an HVAC class or anything so I have no functional vocabulary for these things.
Hugh is sending around a preliminary schematic for the finalized amp, either I or someone else will post it when ready. I saw what I believe to be an error so I am waiting for Hugh's response on that.
- keantoken
Yes, a thick silver colored wire comes through the conduit and is what attaches to the ground post. You'll have to forgive me, I've never been through an HVAC class or anything so I have no functional vocabulary for these things.
Hugh is sending around a preliminary schematic for the finalized amp, either I or someone else will post it when ready. I saw what I believe to be an error so I am waiting for Hugh's response on that.
- keantoken
Hello Hugh
I have at the Haksa schematic you sent me by email, except the need of a 10 ohm ground lifting resistor, all are ok for me.
I need to use parts from my parts bins, so I will use, for the VAS a KSC3503, and for the output MJE15034 and MJE15035.
I have lot of used small and big standard transformers types wall-warts, so I will not need to find a SMPS wall-wart.
Thank
Bye
Gaetan
I have at the Haksa schematic you sent me by email, except the need of a 10 ohm ground lifting resistor, all are ok for me.
I need to use parts from my parts bins, so I will use, for the VAS a KSC3503, and for the output MJE15034 and MJE15035.
I have lot of used small and big standard transformers types wall-warts, so I will not need to find a SMPS wall-wart.
Thank
Bye
Gaetan
Last edited: