Starke Sound Fiera4 4CH Amplifier Does It Meet Power Spec?

[lovinthehd]

The point is he and his fans accused us of the very same thing he apparently does too.

What very same thing are you referring to?

 

[gene]

What very same thing are you referring to?

Thet we held back a test report bc the results were unflattering. We held the report back until we could either test another sample or verify the results with the manufacturer.

I really can't believe this continues on a loop for almost a year.

 

[lovinthehd]

Thet we held back a test report bc the results were unflattering. We held the report back until we could either test another sample or verify the results with the manufacturer.

I really can't believe this continues on a loop for almost a year.

Thanks, just wasn't sure what you were referring to particularly. I'd have to review the timeline but seems mostly Starke was incompetent in this regard in any case.....don't know what they (ASR) did with Starke, but seemed you tried to shield Starke from a particularly stupid release.

ps How do you know that test results were held back just because they were unflattering? Did you discuss that specifically with Amir or is that an assumption?

 

[gene]

Thanks, just wasn't sure what you were referring to particularly. I'd have to review the timeline but seems mostly Starke was incompetent in this regard in any case.....don't know what they (ASR) did with Starke, but seemed you tried to shield Starke from a particularly stupid release.

ps How do you know that test results were held back just because they were unflattering? Did you discuss that specifically with Amir or is that an assumption?

I wasn't shielding anyone. This site was being attacked by Amir and his disciples about our preview article because they couldn't figure out it wasn't a review despite the fact that the article said a review was forthcoming.

Again I'm done rehashing the AMIR and ASR stuff and this thread needs to stay on the topic of the product in question going forward.

 

[lovinthehd]

I wasn't shielding anyone. This site was being attacked by Amir and his disciples about our preview article because they couldn't figure out it wasn't a review despite the fact that the article said a review was forthcoming.

Again I'm done rehashing the AMIR and ASR stuff and this thread needs to stay on the topic of the product in question going forward.

My apologies for pushing on a sore button. Your previews were another issue highlighted by this IIRC.

 

[paulgyro]

I emailed Starke numerous times about it asking for clarification and they never got back to me. So I purchased from another company.

This is a valid concern. I almost purchased the old amp but had questions they didn't answer. I called , emailed, and Facebook Messaged. All went unanswered so I moved on. If they won't help me buy their products how much support will they provide post sale? I'm willing to give them another chance seeing they pulled the old amp from the market.

The new version looks to be a quality product built on a solid Texas Instruments amp chip so worth looking at as a consumer and even more the 8 channel version.

 

[gene]

This is a valid concern. I almost purchased the old amp but had questions they didn't answer. I called , emailed, and Facebook Messaged. All went unanswered so I moved on. If they won't help me buy their products how much support will they provide post sale? I'm willing to give them another chance seeing they pulled the old amp from the market.

The new version looks to be a quality product built on a solid Texas Instruments amp chip so worth looking at as a consumer and even more the 8 channel version.

That's unfortunate they didn't respond to you and hopefully lessons were learned by them from this whole ordeal.

The 8CH version looks like a really good value as well. Looking forward to hearing feedback from users on these new amps.

 

[gene]

My apologies for pushing on a sore button. Your previews were another issue highlighted by this IIRC.

It's all good and thanks for understanding.

 

[paulgyro]

That's unfortunate they didn't respond to you and hopefully lessons were learned by them from this whole ordeal.

The 8CH version looks like a really good value as well. Looking forward to hearing feedback from users on these new amps.

Note the 8 channel is on sale for pre-order @ $1500!

 

[Matthew J Poes]

You have not addressed the poor PCB layout where electrolytic caps are right next to heat-generating components. This for sure will negatively affect reliability, even if those caps are rated at 105°C.

Regarding "the amp doesn’t produce much heat so I suspect in the long run it won’t have a major reliability issue," that is just speculation. Have you measured the amp's efficiency at various power levels and determined the actual heat dissipation in the unvented enclosure?

Regarding "cheaper Chinese amps are a single module per channel with a small 70-100 watt power supply. This has a 1200 watt power supply," a SMPS of 1200 W is a commodity these days -- look at the more complex computer ones to start. Also these "cheaper Chinese amps" already have built-in (as opposed to external brick) SMPS of much more than 70-100 W power (e.g. see the latest review of Aiyima A200 on ASR).

Regarding "each channel is two chips run in BTL [in parallel]," the obvious problem is that due to the very low impedance of the output stage, unless external resistors are employed, there is no reliable way to evenly distribute the load current between the two chips.

Regarding "with a BTL or bridge tied load, you double the voltage differential because the amplifier is run in differential mode. This doubles the power as each half of the amplifier is also now seeing half the impedance," this makes no electrical sense whatsoever. The chip output is differential by design, which means that each of the two output leads "sees" the entire impedance, not half of it. That is because the impedance of the sibling output lead is close to zero and negligibly small compared to that of the load. See schematics starting on p. 22 of the application note. With the same load, a differential output doubles the voltage and quadruples the power, unless a current limit, either in the chip or PS, kicks in.

I am not the designer and it would not be for me to defend their design choices. However the capacitors are in exactly the same place as Ti places them in their own test layout. This amp is just a modification of the original Ti layout.

the Ti chip can be implemented as SE, BTL, or PBTL. Many of the cheaper amps are using the chip in SE, not BTL. In that comparison each amplifier in the chip sees half the load since the speaker ground tied. How does that not make electrical sense. I think you are assuming the amp is always differential. It’s not. That depends on how it’s implemented.

The Topping PA5 is actually the best implementation of this chip. They used TI’s post filter feedback loop, which I am aware of nobody else doing in a production amp using this chip. That amp has a much lower rail voltage and much smaller power supply. It’s under 140 watts. For the same number of channels, the Starke has 600 watts. It’s clearly putting out a lot more power, about double, but with way worse noise. I believe Starke could implement the same feedback loop and have suggested they do so in the future.

As others noted, the PA5 is a not of good quality and as I noted, the Starke is. I would love to see the PA5 implementation scaled up into a proper amplifier as it could be a great option. We shall see if anyone chooses to do so.

 

[Matthew J Poes]

He also publishes far more testing than you do. If he has his own reasons for holding some back, they're his. If you always review/publish what you put on the bench that's nice, but....

I can’t speak for Gene, but I test a ton of stuff we never publish. It’s not Genes decision. I don’t want to waste my time writing up the report. If someone sends us a lemon of a product, we always go to them first with the results. If they stand by the results and claim it’s as expected, we often move forward with the review. However if they find the results surprising and unexpected, I usually wait to see what they want to do. What if they find there is a manufacturing problem that was missed? What if it’s broken and they want to send another sample? What if want to pull the product instantly? I just don’t like wasting my time writing a report for Gene to publish that I know is for a product immediately going off the market.

Stereophile and ASR often publish all measurements and update when new results come in. I understand why, but I think that the first report often stands as the permanent record of a product. If the product was legitimately broken in transit then I see little reason to publish what I know to be not representative results. When I measured the original Starke amp I honestly thought it was broken. The results were so bad I didn’t think it was possible those were representative. I was emailing them waiting to hear that they had found some problem and would send a new sample. When we instead found out it was representative, I figured the record was there already and I just moved on. I actually still have the old amp in my garage.

 

[Matthew J Poes]

It would be great if you can review a couple of those cheap (well, relatively only) chines made amps. At the moment it seems that Amir is the only doing those. So any chance you could spare a few hours to test a couple?

Such as a $99 one:
Amazon.com: S.M.S.L DA-6 Power Amplifier Mini High Resolution DA6 Amp 70W*2 with Remote Control(Silver) : Electronics

DA-6 (smsl-audio.com)

Or that's too cheap to be bothered?

I am very curious about how such a 70 W amp would perform to replace my only integrated amp (NAD C326BE) for my LS50 based system. The 50 W rated NAD has more than enough juice for the LS50 (for desktop use) so if the 70 W rated Chinese amp can do even just half that, i.e. 35 W 4 ohms at 1% THD, it would be good enough for me.

Another one is the PA-5 but over at ASR, it measured well, much better than the Starke one's for sure, but there are quite a few owner's complaints on QA (apparently) related issues that I expected Topping must addressed/solved by now.

PA5 140W x2 Balanced Power Amplifier-TOPPING

The reason we don’t review them is we don’t have contacts with the manufacturers. ASR often has samples sent to them by forum members who own the amp. We don’t review like that. If anyone has contacts with Topping or SMSL I think we would be happy to review their products. I too am very interested in reviewing a PA5. I’ve debated buying one to see, but I have no use for the amplifier. It actually uses the exact chip that is in the Fiera4 but with a better implementation. They appear to be using Ti’s post filter feedback loop. Even if the build quality is poor, it’s fairly cheap. To get .0004% THD+N out of such a cheap amplifier is very impressive.

 

[PENG]

The reason we don’t review them is we don’t have contacts with the manufacturers. ASR often has samples sent to them by forum members who own the amp. We don’t review like that. If anyone has contacts with Topping or SMSL I think we would be happy to review their products. I too am very interested in reviewing a PA5. I’ve debated buying one to see, but I have no use for the amplifier. It actually uses the exact chip that is in the Fiera4 but with a better implementation. They appear to be using Ti’s post filter feedback loop. Even if the build quality is poor, it’s fairly cheap. To get .0004% THD+N out of such a cheap amplifier is very impressive.

Good morning Matthew, really appreciate your direct response with a clear reason and willingness to try testing one such cheap Chines amps. I will let one Topping's staff who I had PM'ed a couple times before and hopefully he could arrange to send you the PA5. I felt sorry for them because some of those folks on ASR were very unforgiving about issues they had, and while I could feed for the owner's frustration, I also felt that one is actually a good example for you to find out via a thorough bench test to show whether like Starke, Topping managed to fix the issues they had with the first (or second, don't know..) to the point it is now worth buying.

You are also the ideal person to do this one as you are very popular and well liked on ASR too. Obviously, we both know that..

 

[lc6]

I am not the designer and it would not be for me to defend their design choices. However the capacitors are in exactly the same place as Ti places them in their own test layout. This amp is just a modification of the original Ti layout.

the Ti chip can be implemented as SE, BTL, or PBTL. Many of the cheaper amps are using the chip in SE, not BTL. In that comparison each amplifier in the chip sees half the load since the speaker ground tied. How does that not make electrical sense. I think you are assuming the amp is always differential. It’s not. That depends on how it’s implemented.

The Topping PA5 is actually the best implementation of this chip. They used TI’s post filter feedback loop, which I am aware of nobody else doing in a production amp using this chip. That amp has a much lower rail voltage and much smaller power supply. It’s under 140 watts. For the same number of channels, the Starke has 600 watts. It’s clearly putting out a lot more power, about double, but with way worse noise. I believe Starke could implement the same feedback loop and have suggested they do so in the future.

As others noted, the PA5 is a not of good quality and as I noted, the Starke is. I would love to see the PA5 implementation scaled up into a proper amplifier as it could be a great option. We shall see if anyone chooses to do so.

TI's PCB layout is just for the chip bench eval kit to be operated in open air, and obviously not to be used in a production model with unventilated enclosure. Following it blindly in a production unit would be a mistake.

You are twisting things. You stated the Starke amp is implemented as BTL and each amp "sees" half the load:
"with a BTL or bridge tied load, you double the voltage differential because the amplifier is run in differential mode. This doubles the power as each half of the amplifier is also now seeing half the impedance. Assuming sufficient current, an 8 ohm load would see double the output. But many then can’t handle 4 ohm loads because each half of the amp sees a 2 ohm load.
This amplifier is already a BTL design."

That is electrical nonsense because both amps in the bridge see the same load, differential voltage doubles and so power quadruples (unless limitations kick in). Similarly, the claim that in a SE design "each amplifier in the chip see half the load since the speaker [is] ground tied" is nonsense. Look at the schematic on p. 26 of the application note. There are two speakers driven by each half of the chip, not one speaker (perhaps with a grounded center tap on a voice coil). So there are four independent channels and each sees the full speaker load, not half the load.

You have still not addressed the issue of heat dissipation in the unvented enclosure.

 

[Matthew J Poes]

TI's PCB layout is just for the chip bench eval kit to be operated in open air, and obviously not to be used in a production model with unventilated enclosure. Following it blindly in a production unit would be a mistake.

You are twisting things. You stated the Starke amp is implemented as BTL and each amp "sees" half the load:
"with a BTL or bridge tied load, you double the voltage differential because the amplifier is run in differential mode. This doubles the power as each half of the amplifier is also now seeing half the impedance. Assuming sufficient current, an 8 ohm load would see double the output. But many then can’t handle 4 ohm loads because each half of the amp sees a 2 ohm load.
This amplifier is already a BTL design."

That is electrical nonsense because both amps in the bridge see the same load, differential voltage doubles and so power quadruples (unless limitations kick in). Similarly, the claim that in a SE design "each amplifier in the chip see half the load since the speaker [is] ground tied" is nonsense. Look at the schematic on p. 26 of the application note. There are two speakers driven by each half of the chip, not one speaker (perhaps with a grounded center tap on a voice coil). So there are four independent channels and each sees the full speaker load, not half the load.

You have still not addressed the issue of heat dissipation in the unvented enclosure.

You are the one twisting what I was trying to say. I didn’t say an SE amp sees half the load.
In a BTL amp, we have two voltage sources and so each voltages source shares the load. Seeing only half of it. That isn’t electrical nonsense as you keep claiming.

 

[lc6]

You are the one twisting what I was trying to say. I didn’t say an SE amp sees half the load.
In a BTL amp, we have two voltage sources and so each voltages source shares the load. Seeing only half of it. That isn’t electrical nonsense as you keep claiming.

Buddy, I do not know which EE school you went to (if at all) and what kind of reading comprehension problems you have (I did not claim you said an SE amp sees half the load, I quoted you saying a BTL amp does). But let me try to explain a bridged amp circuit to you as simply as I can.

You have two (typically, push-pull) amps operating 180 degrees out of phase. When, in the presence of a signal, one produces a voltage, the other produces the opposite voltage, and vice versa (I am purposely not saying positive/negative voltages because in a single-rail power supply, as in the case of TPA3255, both voltages are still positive, with one greater than the other). The amps share a common load, unconnected to the ground, i.e. the load is floating between the two amp outputs and there is no center tap on the load connected to ground. As you said, each amp's output approximates a voltage source, i.e. it has a close to zero output impedance so its output voltage varies very little with load (within power supply and protection limits).

Take the first amp's output as a starting point and, for the sake of a simple example, assume an 8-Ohm resistive load. What does this amp "see" as a load? 8 Ohms (not half of that, or 4 Ohms, as you claim) because at the other end of that load, there is the second amp's output which has close to a zero impedance, as described above.

Also for the sake of example, assume that each amp's output is equivalent to 2.83 V RMS. If each amp had its own 8-Ohm load, it would thus produce 2.83^2 / 8 = 1 W of power. But if the two amps are bridged, they produce a differential voltage of 2 * 2.83 = 5.66 V across the 8-Ohm load, which results in 5.66^2 / 8 = 4 W.

So to recap, in a bridged configuration, each amp sees the entire load (not half of it) and the combined output power is quadruple, not double that in a single-ended configuration (again, assuming that power supply is unchanged and can handle the load, and the internal protection of the amps does not kick in).

 

[Matthew J Poes]

Buddy, I do not know which EE school you went to (if at all) and what kind of reading comprehension problems you have (I did not claim you said an SE amp sees half the load, I quoted you saying a BTL amp does). But let me try to explain a bridged amp circuit to you as simply as I can.

You have two (typically, push-pull) amps operating 180 degrees out of phase. When, in the presence of a signal, one produces a voltage, the other produces the opposite voltage, and vice versa (I am purposely not saying positive/negative voltages because in a single-rail power supply, as in the case of TPA3255, both voltages are still positive, with one greater than the other). The amps share a common load, unconnected to the ground, i.e. the load is floating between the two amp outputs and there is no center tap on the load connected to ground. As you said, each amp's output approximates a voltage source, i.e. it has a close to zero output impedance so its output voltage varies very little with load (within power supply and protection limits).

Take the first amp's output as a starting point and, for the sake of a simple example, assume an 8-Ohm resistive load. What does this amp "see" as a load? 8 Ohms (not half of that, or 4 Ohms, as you claim) because at the other end of that load, there is the second amp's output which has close to a zero impedance, as described above.

Also for the sake of example, assume that each amp's output is equivalent to 2.83 V RMS. If each amp had its own 8-Ohm load, it would thus produce 2.83^2 / 8 = 1 W of power. But if the two amps are bridged, they produce a differential voltage of 2 * 2.83 = 5.66 V across the 8-Ohm load, which results in 5.66^2 / 8 = 4 W.

So to recap, in a bridged configuration, each amp sees the entire load (not half of it) and the combined output power is quadruple, not double that in a single-ended configuration (again, assuming that power supply is unchanged and can handle the load, and the internal protection of the amps does not kick in).

I am confused. Did I say somewhere that you only get double the power? I know it’s 4 times as much if you have sufficient current. I didn’t think I ever said double. If I did, I certainly didn’t mean to say that.

since we have two voltage sources in series, isn’t this the same as saying each sees half the load?

I actually asked Bruno to explain it to me because I wasn’t confident in my knowledge. I went to RIT for EE but did not finish that degree and switched. I’ve been very open about this. I don’t consider myself an expert in electrical engineering by any stretch. Acoustics and psychoacoustics are what I claim expertise. I fumble through electronics and rely on others when I get in over my head.

here was Bruno’s response to me when I shared what I wrote:

“
If you tie the load between a pair of amps running in opposite polarity, you do indeed get twice the voltage (and hence 4 times the power), but only to the extent that the amps have sufficient current capability. By doubling the voltage into the load, you also double the current, after all.

Another way of looking at it is imagining that your load is actually a series connection of two loads of half the impedance. So a 4 ohm load is equivalent to two 2 ohm loads in series. Put that across a pair of bridged amplifiers and the midpoint between the two 2-ohm loads stays at ground potential. Which is as much as saying that each amp has to be able to drive a 2 ohm load, in order for them to be able to drive 4 ohms in BTL.

So now you see that at full tilt, you’re asking each amplifier to deliver 2x as much power compared to the 4 ohm load.”

are we all explaining the same phenomena but disagreeing with exactly how to characterize what is going on? It seems my explanation wasn’t so off from what Bruno said, but I see it also matching what you said.

what am I missing.

 

[gene]

I am confused. Did I say somewhere that you only get double the power? I know it’s 4 times as much if you have sufficient current. I didn’t think I ever said double. If I did, I certainly didn’t mean to say that.

since we have two voltage sources in series, isn’t this the same as saying each sees half the load?

I actually asked Bruno to explain it to me because I wasn’t confident in my knowledge. I went to RIT for EE but did not finish that degree and switched. I’ve been very open about this. I don’t consider myself an expert in electrical engineering by any stretch. Acoustics and psychoacoustics are what I claim expertise. I fumble through electronics and rely on others when I get in over my head.

here was Bruno’s response to me when I shared what I wrote:

“
If you tie the load between a pair of amps running in opposite polarity, you do indeed get twice the voltage (and hence 4 times the power), but only to the extent that the amps have sufficient current capability. By doubling the voltage into the load, you also double the current, after all.

Another way of looking at it is imagining that your load is actually a series connection of two loads of half the impedance. So a 4 ohm load is equivalent to two 2 ohm loads in series. Put that across a pair of bridged amplifiers and the midpoint between the two 2-ohm loads stays at ground potential. Which is as much as saying that each amp has to be able to drive a 2 ohm load, in order for them to be able to drive 4 ohms in BTL.

So now you see that at full tilt, you’re asking each amplifier to deliver 2x as much power compared to the 4 ohm load.”

are we all explaining the same phenomena but disagreeing with exactly how to characterize what is going on? It seems my explanation wasn’t so off from what Bruno said, but I see it also matching what you said.

what am I missing.

I think we are all on the same page here. BTL theoretically doubles voltage which should quadruple power but it rarely ever does due to current limiting. I usually measure 2X the power with a BTL in most products. Virtually EVERY manufacturer that does BTL in consumer audio states the load impedance is 1/2 for each amp in this config and some go as far as to NOT recommending using a BTL with speaker impedances below 8 ohms. This is the case with the NAD M23 I'm about to release my test report on. It simply doesn't have the current drive to double down in BTL for 4 ohm loads.

 

[paulgyro]

Buddy, I do not know which EE school you went to (if at all) and what kind of reading comprehension problems you have (I did not claim you said an SE amp sees half the load, I quoted you saying a BTL amp does). But let me try to explain a bridged amp circuit to you as simply as I can.

Your patronizing and demeaning tone is uncalled for. You think anyone is going care about anything you say after such a snarky opening? Be better then this.

 

[lc6]

I think we are all on the same page here. BTL theoretically doubles voltage which should quadruple power but it rarely ever does due to current limiting. I usually measure 2X the power with a BTL in most products. Virtually EVERY manufacturer that does BTL in consumer audio states the load impedance is 1/2 for each amp in this config and some go as far as to NOT recommending using a BTL with speaker impedances below 8 ohms. This is the case with the NAD M23 I'm about to release my test report on. It simply doesn't have the current drive to double down in BTL for 4 ohm loads.

Of course there are practical limits to a BTL implementation. But this discussion was about theory, in the context of the reviewer's statement that "with a BTL or bridge tied load, you double the voltage differential because the amplifier is run in differential mode. This doubles the power as each half of the amplifier is also now seeing half the impedance." If the statement was something like "this doubles the power supplied by each half of the amplifier (i.e. quadruples the power of the entire amplifier) as each half of the amplifier is also now seeing half the impedance" that would be acceptable, provided a further explanation of the "virtual ground" in the center of the shared load, similar to the one Bruno offered. But, as written, the statement suggested the overall BTL power only doubles (also consistently with practical observations), which is simply not true absent limitations, e.g. for relatively small output levels.

> since we have two voltage sources in series, isn’t this the same as saying each sees half the load?

In a BTL, you do not have two voltage sources in series, you have them at the opposite sides of the shared load. Two voltage sources in series would imply they are directly connected to each other. In that scenario, if one source produced voltage V1 and the other voltage V2, together they would produce voltage V1+V2 and each would see the full (not half of) load because a voltage source has, by definition, zero impedance.

Alright, enough of this already, we beat this horse to death.

Now, since we have @gene 's attention, where is the follow-up to the Yamaha RX-A6A bench test you did last December?
"I will be conducting critical listening tests in the formal review process coming next." Can't find one.