Turbo Switch! For drastically improved quality!
I shall tamper with this! Oh yes, that worked. Faultless audio quality is available at the flip of a switch, but it does cost some output power capacity. For cases when you actually need more output power, just feel free to flip the switch the other way.
This is a manual method, but it certainly is doable.
In our case, a turbo switch would switch the amp between SE or bridge mode.
In SE mode, the gain and output power is much less, in trade for fantastically increased audio quality for relaxed listening levels that we would use most of the time at home. In SE mode it doesn't amplify all the annoying noise from your PC/phone loud enough to hear it. In SE mode the gain is lower, also resulting in lower noise and impressive audio quality. This mode does require output caps, and for audio quality a simple array of two paralleled caps is entirely suitable.
Really, the switch to bridge mode is just there in case of beer and party needs. I'll be trying that directly. In that case, this amplifier's ability for cheap loud concert tickets is suitable.
Here's how I arranged the turbo switch.
I put 6600u (that's 2200u*3) 35v caps with the stripe-side pointed at the power ground.
I put a SPDT switch (3-way switch) with the center pin (the switch constant) going to the speaker jack negative. The other two contacts of the switch, the outboard contacts (the either/or). . . one of them goes to/through the caps for SE mode, and the other pin goes to the negative amp for bridge mode.
In se mode, only one amp output contact is used (that along with the caps series to ground makes up the speaker connection). In bridge mode, at the flip of the switch, then both of the amp outputs are used.
For a stereo implementation, use a DPDT switch so you can switch both channels simultaneously. And in color? But, no, that is just barely beyond the scope of this post.
Instead of my negative/groundside example, the datasheet's positive example can be used just as well. If you need a sketch for either, just mention it. I'm just accustomed to doing it groundside, because that is also compatible with split-rail amplifiers, aka just about universally compatible. I just wrote about what I actually used. I'm enjoying it right now. Basically, what really matters is that you put capacitors in series to the speaker whenever using SE mode. You can put the capacitors series to either speaker contact.
The 6600u is for biggie size speakers; however, a 4" could use 940u (parallel 470u. The paralleling is for quality, a great timesaver) to maximize it and the amplifier by eliminating waste loading (pitches below speaker capacity). Smaller size 4 ohm speakers can use this method and result in higher efficiency, since removing waste loading results in a lighter load (a more powerful system, effectively). Custom size this capacitance to YOUR speaker in order to maximize effectiveness.
Results? This did not work exactly as expected. It didn't run much cooler in SE mode, and that means I need to reduce voltage and/or explore heatsink means. Also, I went hunting the bass booster because it did too much for its own good--Yes, in SE mode, I had to re-eq closer to flat. The apparent output power had not been cut by 4x in SE mode, but rather less than that, perhaps 2/3rds or less. Some datasheet deciphering is needed to figure that one, albeit clues do exist; and, so far, I'm in agreement with the datasheet despite different language to describe matters. Thanks to NXP/Philips for that; and, in that they are uniquely beneficent.
Homework: According to the datasheet, the TDA8932BT that I have, is incompatible with high voltage fun, and I'd actually need the TDA8932BTW for that. Oh darn! Will this maintain its really high end performance in SE mode despite using much safer lower voltage for longevity? I couldn't find a likely ideal 24vdc or 25vdc SMPS, so I'll try 19.5v and report back directly. Let's find out.
I shall tamper with this! Oh yes, that worked. Faultless audio quality is available at the flip of a switch, but it does cost some output power capacity. For cases when you actually need more output power, just feel free to flip the switch the other way.
This is a manual method, but it certainly is doable.
In our case, a turbo switch would switch the amp between SE or bridge mode.
In SE mode, the gain and output power is much less, in trade for fantastically increased audio quality for relaxed listening levels that we would use most of the time at home. In SE mode it doesn't amplify all the annoying noise from your PC/phone loud enough to hear it. In SE mode the gain is lower, also resulting in lower noise and impressive audio quality. This mode does require output caps, and for audio quality a simple array of two paralleled caps is entirely suitable.
Really, the switch to bridge mode is just there in case of beer and party needs. I'll be trying that directly. In that case, this amplifier's ability for cheap loud concert tickets is suitable.
Here's how I arranged the turbo switch.
I put 6600u (that's 2200u*3) 35v caps with the stripe-side pointed at the power ground.
I put a SPDT switch (3-way switch) with the center pin (the switch constant) going to the speaker jack negative. The other two contacts of the switch, the outboard contacts (the either/or). . . one of them goes to/through the caps for SE mode, and the other pin goes to the negative amp for bridge mode.
In se mode, only one amp output contact is used (that along with the caps series to ground makes up the speaker connection). In bridge mode, at the flip of the switch, then both of the amp outputs are used.
For a stereo implementation, use a DPDT switch so you can switch both channels simultaneously. And in color? But, no, that is just barely beyond the scope of this post.
Instead of my negative/groundside example, the datasheet's positive example can be used just as well. If you need a sketch for either, just mention it. I'm just accustomed to doing it groundside, because that is also compatible with split-rail amplifiers, aka just about universally compatible. I just wrote about what I actually used. I'm enjoying it right now. Basically, what really matters is that you put capacitors in series to the speaker whenever using SE mode. You can put the capacitors series to either speaker contact.
The 6600u is for biggie size speakers; however, a 4" could use 940u (parallel 470u. The paralleling is for quality, a great timesaver) to maximize it and the amplifier by eliminating waste loading (pitches below speaker capacity). Smaller size 4 ohm speakers can use this method and result in higher efficiency, since removing waste loading results in a lighter load (a more powerful system, effectively). Custom size this capacitance to YOUR speaker in order to maximize effectiveness.
Results? This did not work exactly as expected. It didn't run much cooler in SE mode, and that means I need to reduce voltage and/or explore heatsink means. Also, I went hunting the bass booster because it did too much for its own good--Yes, in SE mode, I had to re-eq closer to flat. The apparent output power had not been cut by 4x in SE mode, but rather less than that, perhaps 2/3rds or less. Some datasheet deciphering is needed to figure that one, albeit clues do exist; and, so far, I'm in agreement with the datasheet despite different language to describe matters. Thanks to NXP/Philips for that; and, in that they are uniquely beneficent.
Homework: According to the datasheet, the TDA8932BT that I have, is incompatible with high voltage fun, and I'd actually need the TDA8932BTW for that. Oh darn! Will this maintain its really high end performance in SE mode despite using much safer lower voltage for longevity? I couldn't find a likely ideal 24vdc or 25vdc SMPS, so I'll try 19.5v and report back directly. Let's find out.
Well, the Dell 2-wire supply that I had found didn't help audio quality. SO, I plugged in that 3-wire NEC supply, 18.5v used earlier. Apparently, SMPS do vary on quality. It does help to have the sensor lead go all the way to the amp board rather than cut short at the SMPS enclosure. Who knew?
As it turns out, a direct comparison (same supply) as previously used in bridge mode, is even more appropriate. It will be more meaningful. I do so wish I'd have though of it rather than doing things correctly on accident. Gosh.
Anyway, in SE mode, with the same supply, the results are significantly different than bridge mode.
What I was expecting to happen is a whole lot of clipping. I was enjoying the output and didn't change the volume level. Then I lowered the voltage. Theoretically, that should not go well. But, no. Ixnay for making sense. The thing didn't clip at all. I think that the audible quality went up, rather a lot.
For sure, if you're after quality, try single-ended, SE mode. There wasn't any shortfall on quality. None whatsoever.
Problem! I does not boot up in SE mode, as I have implemented it. I had to disconnect the speaker and then re-connect the speaker to get it to play.
Also, I tried the output caps series to positive side. No dice. Then I also had to disconnect and reconnect the speaker before the amp would play. That is not convenient!
So, I also tried the opposite amp. Same behavior. Halp!!!
As it turns out, a direct comparison (same supply) as previously used in bridge mode, is even more appropriate. It will be more meaningful. I do so wish I'd have though of it rather than doing things correctly on accident. Gosh.
Anyway, in SE mode, with the same supply, the results are significantly different than bridge mode.
What I was expecting to happen is a whole lot of clipping. I was enjoying the output and didn't change the volume level. Then I lowered the voltage. Theoretically, that should not go well. But, no. Ixnay for making sense. The thing didn't clip at all. I think that the audible quality went up, rather a lot.
For sure, if you're after quality, try single-ended, SE mode. There wasn't any shortfall on quality. None whatsoever.
Problem! I does not boot up in SE mode, as I have implemented it. I had to disconnect the speaker and then re-connect the speaker to get it to play.
Also, I tried the output caps series to positive side. No dice. Then I also had to disconnect and reconnect the speaker before the amp would play. That is not convenient!
So, I also tried the opposite amp. Same behavior. Halp!!!
Yes that's right, I did notice the caps' 35v rating. I'll be using a LT1085 reg, thinking to push them a little bit lower
The boards may be throwaway cheap, but I wouldn't want failed caps to blow my speakers with DC - can this happen, is the output DC coupled?
Another interesting TDA8932 implementation
I have a couple of these coming - the Taobao price of 35RMB (about $5 with the weakening Yuan) is irresistible. For that money you get a stereo amp with bluetooth receiver in a very nice metal box which feels weighty in your hand (a colleague has already gotten a couple). Another 15RMB nets you the power supply.
https://item.taobao.com/item.htm?spm=a230r.1.14.22.5CD0rj&id=44964882733&ns=1&abbucket=13#detail
Since I have two on order my plan is to monoblock the pair and use them as bass/mid drivers for activating some of my two-way speakers.
I have a couple of these coming - the Taobao price of 35RMB (about $5 with the weakening Yuan) is irresistible. For that money you get a stereo amp with bluetooth receiver in a very nice metal box which feels weighty in your hand (a colleague has already gotten a couple). Another 15RMB nets you the power supply.
https://item.taobao.com/item.htm?spm=a230r.1.14.22.5CD0rj&id=44964882733&ns=1&abbucket=13#detail
Since I have two on order my plan is to monoblock the pair and use them as bass/mid drivers for activating some of my two-way speakers.
Attachments
Don't see the problem here. 34V is within normal specs, these caps normally are 10%+ surge voltage range. But for the expected lifetime, this 5000h is at rated voltage/temperature/ripple - so 35V 105°C. Every 10°C less doubles the lifetime. So running this caps at ~55°C results in:
105°C: 5,000h
95°C: 10,000h
85°C: 20,000h
75°C: 40,000h
65°C: 80,000h
55°C: 160,000h
So having this amp running for 4 hours a day at high temperature/volume/ripple gives you "just" 40,000 days until the caps die. Well, that's over 100 years. If your just need 25 years, you can go up to 75°C. Not enough?
It is not enough.Don't see the problem here. 34V is within normal specs, these caps normally are 10%+ surge voltage range. But for the expected lifetime, this 5000h is at rated voltage/temperature/ripple - so 35V 105°C. Every 10°C less doubles the lifetime. So running this caps at ~55°C results in:
105°C: 5,000h
95°C: 10,000h
85°C: 20,000h
75°C: 40,000h
65°C: 80,000h
55°C: 160,000h
So having this amp running for 4 hours a day at high temperature/volume/ripple gives you "just" 40,000 days until the caps die. Well, that's over 100 years. If your just need 25 years, you can go up to 75°C. Not enough?
The Achilles heel of array decoupling happens if one cap gets different. So, we should not stress them.
Turned up last night. I whipped off the covers and soldered a twisted pair to the power socket. Although I ordered the power supply, it didn't have the matching plug so was useless without hacking. The unit has a 3.5mm audio input which works and the front knob is a digitally controlled volume. I couldn't find the Bluetooth with my mobile phone so I used the wired input. Pleasant enough sound from the phone, no major annoyances so definitely a bargain.
Thanks!
That is the higher resolution implementation, using SE mode. Question: Is it also an TDA8932BTW (more effective to heatsink) version?
In regards to the datasheet: http://www.nxp.com/documents/data_sheet/TDA8932B.pdf
P.S.
Initially finding bluetooth is sometimes just a few seconds window of opportunity after power-cycling the bluetooth device; and so, initial recognizance can take a few rounds. If the phone scans more slowly than the bluetooth device times out (less than a minute for sure), then it may be more like a new hobby.
Last edited:
Thanks for the advice about bluetooth, I did notice there's a blue LED in the antenna which flashes for a while, then stops. I did make a point of pressing 'refresh' on my phone while the flashing was in progress. I have another one which I shall try at some point, curious to hear the effect of sending audio over that kind of link vs the direct wire.
I don't think its the BTW variant - there's a heatsink bar strapped over the top though. Why is the SE mode 'higher resolution' ? I plan to convert mine to BTL via a transformer....
I don't think its the BTW variant - there's a heatsink bar strapped over the top though. Why is the SE mode 'higher resolution' ? I plan to convert mine to BTL via a transformer....
There's lower gain in SE mode.
Specifically, instead of the signal passing through two whole amplifiers with the gain set too high, SE mode has half the trouble because the signal goes through only one amplifier. SE mode has less loading as well, which is more linear and more efficient.
SE mode with 24v actually has enough power for me, so I didn't end up using that Turbo Switch idea to kick it in to BTL mode for approximately three times the power at the cost of halving the quality, because that wasn't needed. I didn't need either the increased power or half the quality, so I didn't need BTL mode.
What the Chinese manufacturers might have missed is the speaker level DC filter cap quality and size. There's two big caps of the same size. They need to be low loss and either 4700u for a big bookshelf speaker or 6800u for large cabinet speaker, and, optionally you can add a 4u7 bypass cap under-board for treble quality (I used a nichicon ES for bypass cap). Other filter sizes are possibly optimal, depending on the speaker, as for example a 4" woofer with a compact enclosure is likely to want a 1000u filter cap, and the board may already have that. What size does that board have? There is an excellent sport to customizing the filter cap size to maximize your speaker, within your application.
If these caps are too big so that the amp doesn't start up, the 47u cap at pin11 (standby delay time) might need to be increased to allow more charge-up time.
Last edited:
Try reversing the order? Press refresh/scan... then power up the bluetooth directly. That'll buy you a few seconds more time. Getting that very first recognizance done can be a workout when both the phone and the bluetooth device have both a tiny and different window of opportunity. One device will probably be slower to link than the other, but I don't know which.
Last edited:
Don't buy this argument - there's the advantage which you forgot to mention of 2nd order distortion cancellation, which may not be very significant.
That's a good point, which can be gotten around by use of a transformer to get a different (i.e. higher) load impedance. I plan trying to trade off the increased power for better quality and see what happens. In other words the trafo at the very least doubles the effective load impedance so the output power is doubled rather than X3 to X4.
Are Chip Amps the End All? Is Everything else obsolete?
New here on this section. Are my eyes deceiving me or is anything else but modular chips amps already obsolete. Sure looks like it to me. Seems like there is some denial of the inevitability of this design in the world. I am floored by the specs I am seeing !!!!!
New here on this section. Are my eyes deceiving me or is anything else but modular chips amps already obsolete. Sure looks like it to me. Seems like there is some denial of the inevitability of this design in the world. I am floored by the specs I am seeing !!!!!
Hi all.
Thanks for the amazingly comprehensive reviews on what seems like a really great little amplifier.
As a complete newbie to the diy audio game, could someone advise me on a good but cheap power supply to match with this amp?
Basically, what is meant by "good quality" power?
What do I look for to find supplies that have "linear regulator chip or older-style warm running SMPS (not the lowest cost new laptop pack)"?
Would something like these do the job, or are switched supplied for LEDs different?
Universal 12 24V 2 3 5 10 15 20 30A Switching Power Supply Driver for LED Strip | eBay
Hotsale AC 85 245V to DC 24V 1A 2A 3A 4A 5A Power Supply Adapter Driver Switch | eBay
I'm looking to build a very cheap system, is a supply for less than e.g. $20 doable?
Another stupidly basic question:
Is the underlying board here the same, with the addition of those handy green cable attachments?
TDA8932 Digital Amplifier Board 15W / 30W / 35W Mono Amplifier Module Low Power | eBay
Thanks very much
Ralph
Thanks for the amazingly comprehensive reviews on what seems like a really great little amplifier.
As a complete newbie to the diy audio game, could someone advise me on a good but cheap power supply to match with this amp?
Basically, what is meant by "good quality" power?
What do I look for to find supplies that have "linear regulator chip or older-style warm running SMPS (not the lowest cost new laptop pack)"?
Would something like these do the job, or are switched supplied for LEDs different?
Universal 12 24V 2 3 5 10 15 20 30A Switching Power Supply Driver for LED Strip | eBay
Hotsale AC 85 245V to DC 24V 1A 2A 3A 4A 5A Power Supply Adapter Driver Switch | eBay
I'm looking to build a very cheap system, is a supply for less than e.g. $20 doable?
Another stupidly basic question:
Is the underlying board here the same, with the addition of those handy green cable attachments?
TDA8932 Digital Amplifier Board 15W / 30W / 35W Mono Amplifier Module Low Power | eBay
Thanks very much
Ralph
In 2016, parts can be a mad chase. That will make LTspice very important (if there's proofed models). I found out the hard way about CSD, Component Specific Design, which is where a crucial part that you relied on goes utterly absent at the last minute, effectively blocking production.
Epic ouch!!!
Today, we regard that sort of thing as bad design.
The current state of the art is a parts tolerant design that allows one to plug in what is available, with only minor resistor value changes (that you figure with LTspice). In addition, today's notion of good design is, from the start, arranged to accept all passable substitutions without fault. Fine tuning after build, is a needful last step, in the past and currently as well.
Good specs are vital, so far as providing an easy path forwards to results. However, the very point when good specs fail is foobar tone or other severe impracticality that is hard to measure. Better measuring continues to be a need. However, several people seem to have managed it, in rather far different ways. It still isn't coming along quickly.
Then as now, make-break points, and the notion of "conditional stability" aka "sweet spot" continue to be important. The amplifier on this thread went against the trend and is excessively stable, providing a fantastic experience with the tone at the expense of imaging/3d/realism. Two workarounds have been found: Enough voltage to destroy it in a far too short timeframe; and, un-bridging it to run it in SE mode more safely. I prefer the latter. Since the gain is factory fixed at too much, the one thing we can do to improve the resolution is to un-bridge it. That worked nicely.
As for specs better than discrete solutions; well, you certainly won't find the slightest bit of evidence here. Instead, here is where you will find doable, and that is quite useful. At that point you'll need to choose something suitable for your application. That is where the fun begins.
Ralph, do you have a 4 ohm or an 8 ohm speaker? Do you have a BTL (bridged) or SE (normal) amplifier?
Those two factors affect appropriate power supply voltage.
The worst is a 4 ohm speaker with a BTL amplifier, which equates to two times a two ohm load. . . which is very hot running.
The best or rather easiest is an 8 ohm speaker with an SE amplifier, representing the least loading. . . which is very cool running.
In those two examples, the maximum voltage that can be used safely, is much different.
To figure this, we'll need to know the amplifier and the speaker.
After that is when we could figure the voltage for your power supply.
The PSRR advantage of running BTL is debatable, despite the figures in the DS. See this article from TI on classD amp measurements :http://www.eetasia.com/STATIC/PDF/200909/EEOL_2009SEP04_ACC_TA_01.pdfhttp://www.whatsbestforum.com/showt...-all-day-long!&p=370819&viewfull=1#post370819