I never said that, just pointing out George speculated then rapidly backtracked saying measurements needed. BTW that audioexpress article mentions Daniel Hertz which always raises a bit of a red flag. But axign are primarily focussed on the PA market from their webpage.
Guess the GAN pushers see PA market as better early adopters for this miniturization and heat reduction technology. For home use i guess MOSFET implementations are small and cool enough and the measurements, compared to the mechanical device they are feeding power, are already overkill. Maybe "atmos 20 speaker amps" could benefit.
But over time I belive GAN will take over consumer market for 100w + amps. But who knows?
There has been some discussion all digital here. If the definition is strictly engeneering it will never be digital as the GAN fets are analog. But if we are more practical and say it is digital if it is not affected by the load (with some limits of cource), we may come so far. But axign clearly don't think we are there yet.
During this discussion i have also learnt that deadtime is affected by the load and that short dead time lowers distortion.
But shortening the deadtime is risky if one do not have control of load. Making mistakes there may result in fire or destroyed speakers. (shoot-through currents). Its a the same with MOSFETS but they provide less strict timing an less maximum current.
My analogy is voltage opamps vs current opamps. Current opamps are "Better" regarding speed and current output but much harder to control and have more catastrophic failure.
But over time I belive GAN will take over consumer market for 100w + amps. But who knows?
There has been some discussion all digital here. If the definition is strictly engeneering it will never be digital as the GAN fets are analog. But if we are more practical and say it is digital if it is not affected by the load (with some limits of cource), we may come so far. But axign clearly don't think we are there yet.
During this discussion i have also learnt that deadtime is affected by the load and that short dead time lowers distortion.
But shortening the deadtime is risky if one do not have control of load. Making mistakes there may result in fire or destroyed speakers. (shoot-through currents). Its a the same with MOSFETS but they provide less strict timing an less maximum current.
My analogy is voltage opamps vs current opamps. Current opamps are "Better" regarding speed and current output but much harder to control and have more catastrophic failure.
Last edited:
Axign clearly state they are targetting the low cost market and getting bang for buck. And many will be happy with a lower heat 1000W amplifier if it has the reliability. Heck most of us over 50 probably would not hear any difference between that and state of the art.
For domestic audio the picture is not clear. TI chips (and others) have taken over the low to mid ground and currently the high ground is a fight between hypex, purifi and icepower. if any of them feel a need to jump to GaN then maybe. I don't know the volumes Orchard ship, but have read good things about their implementation.
At the end of the day if my crystal ball worked I'd not be here, I'd be off winning the lottery 😀
For domestic audio the picture is not clear. TI chips (and others) have taken over the low to mid ground and currently the high ground is a fight between hypex, purifi and icepower. if any of them feel a need to jump to GaN then maybe. I don't know the volumes Orchard ship, but have read good things about their implementation.
At the end of the day if my crystal ball worked I'd not be here, I'd be off winning the lottery 😀
You haven't won the lottery using the magic 8 ball? That's odd. I have won several times over the years. 5 bux here, 2 bux there. But reasonably speaking, don't push it further than that. It is about perspective.
Probably digital is getting better, at least in some ways. I haven't heard the latest, but use it to a degree in my system now by ripping CD's and then I get 2 bux here and there, but nothing way over the top.☕
Probably digital is getting better, at least in some ways. I haven't heard the latest, but use it to a degree in my system now by ripping CD's and then I get 2 bux here and there, but nothing way over the top.☕
Yes, I haven't thought about it before, but as mentioned here, many others has also left the all digital bandwagon.
Se the motor control mixed signal MCUs getting better, so maybe thats the way they think the customer should go? Maybe it's just not enough volume to continue developement of designated chips
It goes to show how bad marketing will often spread like a virus.
Before Class-D amplifiers were getting mainstream, basically everyone knew that PWM isn't digital.
Some main brands started to sell those as "digital", and now it feels a bit like you have to mythbust that dumb claim almost on a daily basis.
I even had EE students coming with this stuff. Sigh......
This is gonna be a copy past from another thread, but it also applies here;
The only benefit could be that the switching frequency could be a little higher.
But that very quickly becomes in the realm of diminishing returns.
Standard MOSFET technology is fine switching around 400-1000kHz.
Maybe in the ultra high power applications (> 2kW) it might be helpful with power dissipation.
Little addition for that last part;
Even with just MOSFETs we are able to get to about 95% efficiency (if not a little more).
So in total, there isn't an awful lot to win here.
As for distortion, the biggest losses are very often in the inductor, not in the main switching circuit itself.
Only a proper post-feedback solution will fix that.
I have seen MANY designs with the wrong kind of inductors being used, as well for the material as well as the size.
Me neither, also the smaller part argument doesn't make sense, since the bulk of any PCB is being taken by other things like capacitors, connectors, controller and inductors/coils.
The only benefit could be that the switching frequency could be a little higher.
But that very quickly becomes in the realm of diminishing returns.
Standard MOSFET technology is fine switching around 400-1000kHz.
Maybe in the ultra high power applications (> 2kW) it might be helpful with power dissipation.
Little addition for that last part;
Even with just MOSFETs we are able to get to about 95% efficiency (if not a little more).
So in total, there isn't an awful lot to win here.
As for distortion, the biggest losses are very often in the inductor, not in the main switching circuit itself.
Only a proper post-feedback solution will fix that.
I have seen MANY designs with the wrong kind of inductors being used, as well for the material as well as the size.
A much more important question would be;
IF this would be all true, why on EARTH would you just make yet another amplifier as a company?
And not patent your new idea and license it out.
Because it would mean that the rest of the world has been sleeping for the last 3 decades if not more.
^ I know your question was rhetorical.
Threads like this are informative.
I'm a newcomer to DIY, but not to audio as a hobby. The cynic in me is beginning to believe that the engineering and marketing teams at many audio companies are joined at the hip. I'll give them credit for being savvy, but I abhor 'them' for (I think) knowingly taking advantage of three things. 1) Loose consumer protection / fraud laws and enforcement as they relate to marketing claims (at least in the USA). 2) The ignorance of their consumer base, including me. 3) The multitude of complexities and arguments surrounding what measures better and what sounds better, particularly when the nearly limitless combinations of additional gear and room acoustics come into play. I'm still learning, but I've come to believe that an amplifier 'should' be the easiest of all audio gear to measure to see if it's behaving properly.
We're easy pickings for people designing amplifiers. An amplifier has but one engineering purpose, but limitless consumer expectations. I know it's a tad more complex. However, my point is: I think most amplifiers used to really have sonic character that could be distinguished simply because the tech wasn't there to create "perfect amplifiers". Thus, in the not-too-distant past two "really good" amplifiers could actually behave differently from a measurement and audibility standpoint. I guess a lot, but I'll bet that in years past the audio engineers would have beaten each other with sticks to achieve the performance available today. Their design goals were likely based on distortion characteristics and power handling. Since those problems have been 'solved', today's marketers have taken advantage and created their own limitless consumer expectations around sonics. Thus, they can 'redesign' iteratively. My guess is that you'll start to see things like "New GaN technology that sounds like tubes." Why? OK... I think I know why. Sales. With the exception of price... you can't just say, we made the same thing again. They have to keep us feeding from the trough of 'improvement'... Since engineering goals have been met, those expectations have to come from the ever-debatable sonic improvements. If they didn't, then the market would have dried up 20 years ago. Yet, we see new companies cropping into the mix all the time, and they're not marketing to industrial or commercial users. They're marketing to 'audiophiles'. I'm not in any way implying that there are not some great engineering achievements as it relates to consumer audio. I just think that for the largest extent, it's irrelevant to most consumers.
In fairness, I am perfectly content to build effects boxes and see what I happen to like. I also really like to solder stuff. I'm still trying to find a "great" Class-D or another chip-amp design that has exceptional measurements from whatever reviewer or forum to try. I've bitten the apple from the "old-school", and I want to see what some of this new-fangled stuff is all about.
Thanks to all in the thread for shedding a little light on how some of this works and what may "matter" about the design and engineering.
Threads like this are informative.
I'm a newcomer to DIY, but not to audio as a hobby. The cynic in me is beginning to believe that the engineering and marketing teams at many audio companies are joined at the hip. I'll give them credit for being savvy, but I abhor 'them' for (I think) knowingly taking advantage of three things. 1) Loose consumer protection / fraud laws and enforcement as they relate to marketing claims (at least in the USA). 2) The ignorance of their consumer base, including me. 3) The multitude of complexities and arguments surrounding what measures better and what sounds better, particularly when the nearly limitless combinations of additional gear and room acoustics come into play. I'm still learning, but I've come to believe that an amplifier 'should' be the easiest of all audio gear to measure to see if it's behaving properly.
We're easy pickings for people designing amplifiers. An amplifier has but one engineering purpose, but limitless consumer expectations. I know it's a tad more complex. However, my point is: I think most amplifiers used to really have sonic character that could be distinguished simply because the tech wasn't there to create "perfect amplifiers". Thus, in the not-too-distant past two "really good" amplifiers could actually behave differently from a measurement and audibility standpoint. I guess a lot, but I'll bet that in years past the audio engineers would have beaten each other with sticks to achieve the performance available today. Their design goals were likely based on distortion characteristics and power handling. Since those problems have been 'solved', today's marketers have taken advantage and created their own limitless consumer expectations around sonics. Thus, they can 'redesign' iteratively. My guess is that you'll start to see things like "New GaN technology that sounds like tubes." Why? OK... I think I know why. Sales. With the exception of price... you can't just say, we made the same thing again. They have to keep us feeding from the trough of 'improvement'... Since engineering goals have been met, those expectations have to come from the ever-debatable sonic improvements. If they didn't, then the market would have dried up 20 years ago. Yet, we see new companies cropping into the mix all the time, and they're not marketing to industrial or commercial users. They're marketing to 'audiophiles'. I'm not in any way implying that there are not some great engineering achievements as it relates to consumer audio. I just think that for the largest extent, it's irrelevant to most consumers.
In fairness, I am perfectly content to build effects boxes and see what I happen to like. I also really like to solder stuff. I'm still trying to find a "great" Class-D or another chip-amp design that has exceptional measurements from whatever reviewer or forum to try. I've bitten the apple from the "old-school", and I want to see what some of this new-fangled stuff is all about.
Thanks to all in the thread for shedding a little light on how some of this works and what may "matter" about the design and engineering.
Last edited:
Good points. I also like to mention that I've heard stories from engineering friends who were 'requested' to cut some corners without compromising performance to lower the cost. Or, if that can't be done, just compromise performance.
I recently got into an argument with someone who told me that 'you can't manufacture and sell a top of the line DAC for under $1k'.
My point is that you can, if you build and sell 100,000 units. You cannot if you build and sell 50 per year. It's that simple.
If you sell two or three amplifiers per year, they have to retail at $ 50k, no matter what you put into them or how good or bad they are.
Jan
I recently got into an argument with someone who told me that 'you can't manufacture and sell a top of the line DAC for under $1k'.
My point is that you can, if you build and sell 100,000 units. You cannot if you build and sell 50 per year. It's that simple.
If you sell two or three amplifiers per year, they have to retail at $ 50k, no matter what you put into them or how good or bad they are.
Jan
I use premium flat copper wire high current inductors in my output filters and it makes a difference. These four inductors alone are the most significant cost of the amp assembly.
@jan.didden you could equally argue that audiophiles and the hifi press won't take you seriously if you try and sell a top of the line DAC for under $1000 🙂
^ @jan.didden
I promise I'm not thread stalking. I just find the threads you're involved in to be informative.
My 'past life' has made me a cynic, sadly. I can't say that I've "seen it all", but I've seen a lot. The funny thing to me is that in my past, marketing and R&D / Engineering were often at odds. Engineering wouldn't want to compromise. Sometimes for good reasons, and sometimes for simple ego.
Good - the chances of this causing harm increase at X rate going from this part to that part.
Ego - I forgot to design into an expected cost... I designed the awesomest of awesome, and I will not for any reason make it less awesome. Who cares if the market won't bear a 1000% increase in cost over the competition for a 1% improvement? I made the best. MINE IS NOW SotA.
Your dilemma seems to be caught somewhere in the middle. With that said, I don't know what top of the line means in your terms. Some might have it to mean the fanciest of chassis with a few ready-made parts available from common suppliers. You might actually mean engineering performance. Working with sales teams to understand customers and what they may want and pay for ... is fun.
The best job I ever had was in R&D, but I reported to the lead of sales. It sounds wonky, but we had a fantastic relationship.
Sadly with some companies, when the marketer and a lead engineer are one and the same or are running blissfully hand-in-hand... it raises the hairs on my neck depending on the marketing claims.
I hope... I truly hope... that I'm very, very wrong. Perhaps I haven't learned enough yet, perhaps there is a level of deceit and perhaps there's a level of incompetence. I think it's a combination of the three and maybe a few other things.
Still, I'm trying to be less of a cynic. It's overused, and often out of context - "Trust but verify".
I promise I'm not thread stalking. I just find the threads you're involved in to be informative.
My 'past life' has made me a cynic, sadly. I can't say that I've "seen it all", but I've seen a lot. The funny thing to me is that in my past, marketing and R&D / Engineering were often at odds. Engineering wouldn't want to compromise. Sometimes for good reasons, and sometimes for simple ego.
Good - the chances of this causing harm increase at X rate going from this part to that part.
Ego - I forgot to design into an expected cost... I designed the awesomest of awesome, and I will not for any reason make it less awesome. Who cares if the market won't bear a 1000% increase in cost over the competition for a 1% improvement? I made the best. MINE IS NOW SotA.
Your dilemma seems to be caught somewhere in the middle. With that said, I don't know what top of the line means in your terms. Some might have it to mean the fanciest of chassis with a few ready-made parts available from common suppliers. You might actually mean engineering performance. Working with sales teams to understand customers and what they may want and pay for ... is fun.
The best job I ever had was in R&D, but I reported to the lead of sales. It sounds wonky, but we had a fantastic relationship.
Sadly with some companies, when the marketer and a lead engineer are one and the same or are running blissfully hand-in-hand... it raises the hairs on my neck depending on the marketing claims.
I hope... I truly hope... that I'm very, very wrong. Perhaps I haven't learned enough yet, perhaps there is a level of deceit and perhaps there's a level of incompetence. I think it's a combination of the three and maybe a few other things.
Still, I'm trying to be less of a cynic. It's overused, and often out of context - "Trust but verify".
Last edited:
A well run company will tell the Engineer what the BOM cost should be for a product before he starts designing as you don't start product design without a target market. this isn't always right. For example you hear of cases where a product doesn't sell at the initial price but sells when a go faster stripe is put on it and the cost doubled (but these are rare). A company where the engineering dept are given free reign to design the 'best' cost no object are in the minority and often small hifi companies or other 'lifestyle' manufacturers.
^ 
!!!!
One of the great benefits (for me) of reporting to sales was learning their often well-informed thoughts on how competition was performing in terms of relative sales. I (and many) engineers may have overlooked some things customers will pay for. Before I worked for him, I was all about 'the better product wins'. Oh, how wrong I was. In an equal and opposite manner, we could tell the sales team where we thought we had clearly outperformed in areas we thought mattered. They may choose to then 'tell the customer more clearly' that ours was better in that area, thus creating a marketing point. It was fun.
I had a great relationship with that boss b/c I came from one of our customers. I moved "down the supply chain". So, I could sit in on customer sales meetings, listen carefully to what was said in sales and marketing meetings with clients. It shortened the design cycle considerably. I spoke the customers' engineering language / and knew their manufacturing processes. So, I was allowed to talk directly to their engineers also as needed. It's a rare thing, but I could work with my boss to determine costing or cost averaging. If they want 'specialty product X', sure ... We can give them exactly what we've got in inventory. You put a new name on it, and I'll work with their engineers on integrating it into their existing processes. You charge whatever they'll pay. I'm staying out of that. I cut the number of customized-to-client solutions 10-fold.
It's important to note that I've only worked for one company that sold directly to consumers. That was an entirely different experience.
!!!!One of the great benefits (for me) of reporting to sales was learning their often well-informed thoughts on how competition was performing in terms of relative sales. I (and many) engineers may have overlooked some things customers will pay for. Before I worked for him, I was all about 'the better product wins'. Oh, how wrong I was. In an equal and opposite manner, we could tell the sales team where we thought we had clearly outperformed in areas we thought mattered. They may choose to then 'tell the customer more clearly' that ours was better in that area, thus creating a marketing point. It was fun.
I had a great relationship with that boss b/c I came from one of our customers. I moved "down the supply chain". So, I could sit in on customer sales meetings, listen carefully to what was said in sales and marketing meetings with clients. It shortened the design cycle considerably. I spoke the customers' engineering language / and knew their manufacturing processes. So, I was allowed to talk directly to their engineers also as needed. It's a rare thing, but I could work with my boss to determine costing or cost averaging. If they want 'specialty product X', sure ... We can give them exactly what we've got in inventory. You put a new name on it, and I'll work with their engineers on integrating it into their existing processes. You charge whatever they'll pay. I'm staying out of that. I cut the number of customized-to-client solutions 10-fold.
It's important to note that I've only worked for one company that sold directly to consumers. That was an entirely different experience.
^ 'Design to manufacture' is... often overlooked. The nice thing about a small company was that that same engineers were in the lab and "on the line", but you're dead on IMO. I've never experienced it directly, but I've heard stories of lab performance never matching manufacturing performance and/or hand-built assemblies that couldn't be reasonably mass-produced to a cost target using tooling available or reasonable labor cost.
It seems so 'common sense' now... but... oof.
Anyway... I think I've moved this away from the original topic. Apologies. I'll sit back and "listen" now. Looking forward to learning more about "digital" amps and this GaN stuff.
It seems so 'common sense' now... but... oof.
Anyway... I think I've moved this away from the original topic. Apologies. I'll sit back and "listen" now. Looking forward to learning more about "digital" amps and this GaN stuff.
- Status
- Not open for further replies.