7498E BTL Mode Question

Hello,

I have bought a couple of these amps with the intention of each one powering a passive subwoofer. I decided to use them as mono amps for this purpose.




I have a couple of questions regarding BTL mode.

I have close the jumpers on J1, J2, J3 and J4 as below:




I am running each board off a single 33V 10 Amp power supply

When in BTL mode should the output only come out of one of the L&R speaker outputs on the board? I seem to get an output on both. I assume it doesn't matter which one I connect to.
Does BTL mode increase the power output for one driven channel - I can't seem to notice to much of a difference from running the same speaker off a stereo channel?

Also, I notice that when not in BTL mode, the fan runs all the time. When in BTL mode in doesn't seem to power on even when driven hard.

Thanks
 
Last edited by a moderator:
Situation very well explained, much appreciated.
As Chris has already explained, the TDA7498E chip is conceived as "BTL" using two outputs operating in counter-phase. On the rear side of your boards it is stated that you have to connect these 4 jumpers and you will have "BTL". It is wrong, you already have "BTL".

If you look in the TDA7498E datasheet, they describe how to connect for "PBTL". This is actually what the jumpers do.
"PBTL" does not increase the output voltages to the speakers and with the same speakers you notice no difference at low outputs levels or with 8 Ohm speakers. What does "PBTL" then do? - put the two set of output switches in parallel such that the amplifier can handle low impedance loads and give more output power through increased current capability (for the low impedance loads).
The datasheet states the PBTL output power with a 3 Ohm load (220W/170W). The chip can handle two 4 Ohm loads in normal stereo/BTL coupling and with 2x160W/125W stereo output power. So, the gain in output power is not enormous by going to PBTL and a single 3 Ohm load.

Now you anyway have bought the two boards, I would still use them in mono/PBTL coupling simply because your output switch impedance is halved and the chip will have less power loss.

You are right, you can use any of the two outputs. You just connect the chip output switches to one another (J1 and J2) but do not reconfigure the output filters. Therefore, both outputs are available. You have to try which input is active.
 
Last edited:
@Chermann - Thanks for pointing this out. I wasn't aware about the chip already being in BTL mode.

@FauxFrench - Thanks for the deep information - very interesting, especially about how it uses PBTL for low impedance loads. You seem to be very knowledgeable about these little amps! Like you say, I think I will keep them in this configuration as I have two.

Thanks guys!
 
@Chermann - Thanks for pointing this out. I wasn't aware about the chip already being in BTL mode.

@FauxFrench - Thanks for the deep information - very interesting, especially about how it uses PBTL for low impedance loads. You seem to be very knowledgeable about these little amps! Like you say, I think I will keep them in this configuration as I have two.

Thanks guys!


as i said..Fauxfrench is the expert.;)


please feed us with your mods and pics...and measurements and SQ.
:D
 
@Chermann - Thanks for pointing this out. I wasn't aware about the chip already being in BTL mode.

@FauxFrench - Thanks for the deep information - very interesting, especially about how it uses PBTL for low impedance loads. You seem to be very knowledgeable about these little amps! Like you say, I think I will keep them in this configuration as I have two.

Thanks guys!

Dalgairns,

I forgot one detail.
You should preferably remove the two filter chokes in the output you do not use. You have probably read that a class D amplifier output should not be unloaded with the amplifier turned ON. You have an output you do not use and you can only disable that output filter by removing the two filter chokes.
 

ICG

Disabled Account
Joined 2007
I have a couple of questions regarding BTL mode.

[...]

I am running each board off a single 33V 10 Amp power supply

Well, with that configuration you maybe have to accept a heavy loss on maximum power because (as already explained) the maximum power is only achieved at much lower impedances (3 Ohm) and (this was not mentioned before though) only at 36V. The important question is now: What impedance are you driving? If you are using 4 Ohm drivers, I would not change the configuration. If you are using 8 Ohm subs and the 33V PS though, according to the datasheet, you're looking at around a meager ~70W @10%THD. That's especally disappointing because the TDA7498 (non-'E') would deliver ~78W with the same setup on each channel which means it would have been the better choice there. The TDA7498E got a maximum Vcc voltage of 45V. If you'd increase the power supply voltage to 44V, the power output would be around 130W@8Ohm, almost the double. So, if you are using 8 Ohm drivers, a higher PS voltage is much preferred but you might have to change some parts on the amplifier board, likely the buffer capacitors, maybe more.

I forgot one detail.
You should preferably remove the two filter chokes in the output you do not use. You have probably read that a class D amplifier output should not be unloaded with the amplifier turned ON. You have an output you do not use and you can only disable that output filter by removing the two filter chokes.

That is definitely WRONG! Firstly, if you remove the chokes, you won't have any output anymore since the two channels are already bridged. Removing the two chokes results in disconnecting the output. (See datasheet ch. 3.4, page 10) If you want to eliminate the 2nd choke in PBTL mode, you'd have to replace it with a bridge. Secondly, the second choke is needed for the low output impedance, otherwise the amp will show a significant treble increase because of the interaction of the speaker impedance with the changed output filter - which is likely not the case here because: Thirdly, that is irrelevant because he intended the amp for subwoofers, in that range the output filter does next to nothing (no matter in which configuration) which makes these changes completely irrelevant/unneccessary.
 
That is definitely WRONG! Firstly, if you remove the chokes, you won't have any output anymore since the two channels are already bridged. Removing the two chokes results in disconnecting the output. (See datasheet ch. 3.4, page 10) If you want to eliminate the 2nd choke in PBTL mode, you'd have to replace it with a bridge. Secondly, the second choke is needed for the low output impedance, otherwise the amp will show a significant treble increase because of the interaction of the speaker impedance with the changed output filter - which is likely not the case here because: Thirdly, that is irrelevant because he intended the amp for subwoofers, in that range the output filter does next to nothing (no matter in which configuration) which makes these changes completely irrelevant/unneccessary.

I assume we discuss the situation where we configure the chip to run with all four outputs synchronous (PBTL), in two counter-phased pairs, such that the chip outputs can be connected with jumpers (J1&J2).

Fig.3 on page 10 of the TDA7498E datasheet (ST) divulges the initial stereo situation (BTL) which then has to be modified into PBTL by connecting the chip output switches in parallel. If we do that on the circuit of Fig.3, we end up with two counter-phased outputs looking into two sets of output chokes (L1-L2 / L3-L4) each being connected to two speaker outputs that remain from the original stereo (BTL) configuration.
As the output from the chip is now a single channel (mono), we can choose either of the speaker outputs as they have the same output signal. Say, we choose the output through L1 and L2, L3 and L4 are not used. That will be a series resonant circuit (L3, L4, C26) without damping and without use.

What purpose do you see from maintaining that obsolete output filter and speaker output?
 

ICG

Disabled Account
Joined 2007
Fig.3 on page 10 of the TDA7498E datasheet (ST) divulges the initial stereo situation (BTL) which then has to be modified into PBTL by connecting the chip output switches in parallel. If we do that on the circuit of Fig.3, we end up with two counter-phased outputs looking into two sets of output chokes (L1-L2 / L3-L4) each being connected to two speaker outputs that remain from the original stereo (BTL) configuration.
As the output from the chip is now a single channel (mono), we can choose either of the speaker outputs as they have the same output signal. Say, we choose the output through L1 and L2, L3 and L4 are not used. That will be a series resonant circuit (L3, L4, C26) without damping and without use.

What purpose do you see from maintaining that obsolete output filter and speaker output?

There are two scenarios: First, you parallel the channels before the filters and bridge the resulting output. Then only one filter will be used (per side). But then it's useless to remove the choke because it does not do anything anymore anyway - but the remaining filter has to take the double current and the saturation occurs much earlier and the distortion rises vastly. So why do the work with no improvement (or even the opposite)?

Second, you bridge the channels and parallel the channels after the filters. There the removal of the filters would result in uneven behaviour and possible cross-currents and parts of the amp trying to correct the others. Aside from that, on half of the amp would be much more sensitive to HF interference.

Either way, it's not favorable. The only improvement would be to replace the chokes for better/bigger ones.
 
There are two scenarios: First, you parallel the channels before the filters and bridge the resulting output. Then only one filter will be used (per side). But then it's useless to remove the choke because it does not do anything anymore anyway - but the remaining filter has to take the double current and the saturation occurs much earlier and the distortion rises vastly. So why do the work with no improvement (or even the opposite)?

Second, you bridge the channels and parallel the channels after the filters. There the removal of the filters would result in uneven behaviour and possible cross-currents and parts of the amp trying to correct the others. Aside from that, on half of the amp would be much more sensitive to HF interference.

Either way, it's not favorable. The only improvement would be to replace the chokes for better/bigger ones.

First scenario: The output filter is designed to handle 4 or 8 Ohm loads in stereo mode. Looking at the size of the chokes, it seems likely they can handle 10A. If the configuration is changed to PBTL, the current in the chokes (one filter only) does not change unless the load is reduced (to 3 Ohm). Speakers with 3 Ohm nominal impedance are not that frequent so it is likely to be 4 Ohm. We agree that removing one set of chokes (from the filter not being used) does not leave a situation without output. Then, we are in the situation described by "ddapkus" for TPA3116 with an output filter not being loaded - the resonance may be exited and damage caused. This is why I suggest to remove the two unused chokes.

Second scenario: There would be no need for jumpers on the board as the connections could just be made at the output terminals. I fully agree with you that such uncontrolled paralleling of outputs may leave poor load-sharing and worse, cross-currents. With the jumpers to be mounted on the board, this is not what the designers envisaged and it is strongly discouraged. Such paralleling would not be PBTL as described in the datasheet.

I also agree, had the designers initially made the board for PBTL (mono) operation, they would have used two filter chokes only and perhaps chosen chokes with a higher current rating.
 
Last edited:

ICG

Disabled Account
Joined 2007
First scenario: The output filter is designed to handle 4 or 8 Ohm loads in stereo mode. Looking at the size of the chokes, it seems likely they can handle 10A. If the configuration is changed to PBTL, the current in the chokes (one filter only) does not change unless the load is reduced (to 3 Ohm). Speakers with 3 Ohm nominal impedance are not that frequent so it is likely to be 4 Ohm. We agree that removing one set of chokes (from the filter not being used) does not leave a situation without output. Then, we are in the situation described by "ddapkus" for TPA3116 with an output filter not being loaded - the resonance may be exited and damage caused. This is why I suggest to remove the two unused chokes.

No, that's not the scenario, it is not what happened with the TPA3116. There the output of the amp was completely without load, not the filter 'dangling freely' while the output still under load. That are two completely different things, that is not happening here, the amp output still sees a load. But if you'd want to prevent that from happening, it would be a much easier to change and reversible thing to cut the lead to the capacitor(s). That way you could reconnect it with a simple soldering point or a jumper and you don't have a chance to lose the choke.

Second scenario: There would be no need for jumpers on the board as the connections could just be made at the output terminals.

It would be a very bad design choice doing it without jumpers because if you leave them out both channels would run asynchronously (you have to synchronize the clock!) and to be honest, I don't know where/what these jumpers exactly connect and I assume neither do you. I assume though, they did design the amp that way if you are PBTL'ing it, it will not blow up without removing chokes.

I also agree, had the designers initially made the board for PBTL (mono) operation, they would have used two filter chokes only and perhaps chosen chokes with a higher current rating.

The TDA7498E amps are generally a bad choice for PBTL use, you can easily see the power output figures in PBTL are so much lower than both channels combined (instead of 160+160=320W just only 220W). Why is the power so much lower in PBTL? The TDA7498E cannot handle the high current. I've read about and also seen myself TDA7498E amps wich were used near the power limit, becomming that hot they liquified the solder on the pins and that lead to a short circuit, frying chip & board.
 
Interesting discussion!
Especially as I have this red Sanwu 7498E and the blue board Feixang board and I am also confused about how to run the red board in PBTL.
I found some nice burned examples online. :devilr: But that might just be from overheating as ICG pointed out.

The Feizang/blue board has solder points/blobs to connect a PBTL/mono load, right?
While looking at it I noticed, 3 pins with "mode" written above them in the upper right corner. Does anyone has a clue what these are for?
Switching BTL - PBTL? Or maybe switching between the differential and the RGL input?

I'm thinking of driving some PA tweeters with it, fed by a 3 way digital crossover, so not much power. (hope to start this project soon)

These are my own pics as I could not find a pic of the bottom side of the blue board online.

An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.
 
Last edited:
I only have the red Sanwy board and have never used it in PBTL-mode or even near its limits in normal BTL operation. For BTL operation up to 2x100W, I find it is a fine chip but using it for currents above 10A in a 200W amplifier seems risky. I'm not surprised if some TDA7498E chips do not manage. It is a small PowerSSO-36 housing and, though they use double chip-pins for power signals, also the PCB tracks become very tiny. And, the thermal contact with the heatsink becomes delicate.

Our discussion was on a very restricted basis: what to do when you have already bought two Sanwu boards to drive two sub-woofers? On the blue board you have, I have no information.

The TDA8954 is another example of a rather small SMD-housing (HSOP24) claiming use for a 420W (!) amplifier. The TDA8954 can also be supplied in a DBS23P housing (DIL-type) and perhaps this housing is implied for the 420W use.

SMD-housings are evidently used to offer the possibility of robot-mounting but from an engineering point of view, I cannot see so thermally marginal designs as appropriate. The (very) small chip-housings anyway require considerable heatsinks so why not give the chips better cooling pads and decent chip-leads (pins)?

But, as usual they don't ask me before deciding on a chip-housing and let the marketing people exaggerate the performance ;-))
 
Last edited:
Hello,I have burnt one chip driving a dual voice coil Pioneer 4th order band pass sub .Each voice coil 2,7ohm. Used the red bord with fan but mounted it bad,did not get enough cooling and played LOUD. So I ordered 2 new board and thought that one for each voice coil,that would make them last. On the board I se that left is mono in but output? Use left or right or parallel them,left and right + togehter,same with - ?
 
If you use the red board, configured for PBTL, you can use either of the outputs (left or right). As explained above, I would remove the filter chokes that are not used but "ICG" disagrees with that (see his argumentation above).
If you configure the red board for PBTL and use the two outputs in parallel, you effectively reduce the filter inductance to half while the filter capacitance becomes double. I do not know how the chip will react on that - I wouldn't do it.

I notice you are going to use the amplifiers with a low impedance load. Then, do only supply the amplifiers with a voltage that does not allow the output currents to be excessive. All here agree, cooling of the chip and the loading of the circuit tracks and pins is delicate with such a small high-power chip. These TDA7498E chips are know to have died from high loading.
 
Last edited:
Hello and thanks. I used 30volts when it died at high volume. The bandpass sub is very sensitive so I may not need dual amps. I ordered 2 amps cause I thought it would ease the load on the amp(s). But maybe it doesn't since it is the same single chip that gets the load,am I right or wrong? I have a couple of 2ohm 50watt resistors. I could put them in series after the amp to ease the load. Parallell 2resistor to get 1 ohm or would you use 2ohm. And what voltage would be safe with or without resistors? Many maybe stupid questions but I would make it last,I sometimes play loud in my car :) I have adjustable stepup so almost any voltage is possible..
Filip
 
Hey. I connected one coil to left and the other to right. So it was 2,7r per coil. That's why i think maybe 1 or 2 ohm in series so it will be 3,7 or 4,7ohm if I use one amp i stereo. I only afraid it will soften the punch in the bass. Would voicecoil in series have the same effect? I now use another TDA7498, not E-version. And use 2 ohm in series to avoid burning it up. Tried with 4ohm but that was bad punchless bass. It's not so good now either but it's the lower power 7498 also. But you got a good point about serieconnection,thanks

Filip
 
I will support the suggestion of Think: connect the two voice-coils in series and you have a slight reduction of output power, no resistors just dropping energy and less current stress in the amplifier.

I do not really understand your 1/2 amplifier question.
Normal BTL (stereo) with one speaker at each output - the same (one) chip is used.
PBTL operation with one speaker per amplifier - two chips are used.
PBTL operation with two speakers connected in series at one amplifier - the same (one) chip is used.