Electronics > Projects, Designs, and Technical Stuff

Texas Instruments 3116D (ClassD) amp chip

<< < (9/10) > >>

DW1961:

--- Quote from: cv007 on August 18, 2020, 06:56:29 am --->In this case, I would assume the relationship between figures 14 and 22 could also be applied to figure 13, that is: at 24V, power at 8 ohms in PBTL would be double when compared to BTL

You are missing the fact that the PBTL graph is using 2 Ohms. All the graphs essentially end up the same (BTL 8,4,PBTL 2), each is just changing the load impedance to increase the power. PBTL is already BTL, its just parallel to share the heat.

If you have 24v and 8 Ohms, it makes no difference if you run in BTL or PBTL as your total output will be the same. What you get, as I have said previously, is the ability to direct all the power into a single low impedance speaker if wanted. With 8 ohms and PBTL, all you get is less heat for each ic, which is a good enough reason and could be the main reason they run them in PBTL.

--- End quote ---

Spot on for resistance. The only way an amp "makes more power" is either getting more voltage from the supply or lowering resistance so more current can flow to it. At least that's how I am understanding it currently.

"If you have 24v and 8 Ohms, it makes no difference if you run in BTL or PBTL as your total output will be the same." Clear, and thank you.

What I have been saying all  along in this thread in one way or another, which is why I created the thread. I just didn't see the power benefit of using two chips vs one given the same voltage restrictions.

"all you get is less heat for each ic, which is a good enough reason and could be the main reason they run them in PBTL." ---Yo!, thanks.

However, the single amp chip says it can run full out without any problems heat wise. If they really wanted to reduce chip heat, it seems like they cold do it a lot cheaper buy using a bigger heat sink or active cooling. Fans and heat sinks are a lot cheaper than the 3116D chip, I would assume, and less engineering too.

The seller on Amazon wrote me back and said the amp package they offer with x2 3116D chips is 24V and anything higher will burn it. He also said that using the included power supply at 24V, you're gonna get 70 watts continuous power at 8 ohms. That's the same as a single chip, according to the graph on the spec sheet we've been referring to.
 

I keep asking them why use 2 chips then? We'll see if they ever answer.

tooki:
You’re not going to get a reply from an engineer. Expecting a reply like that is an exercise in frustration.

DW1961:

--- Quote from: tooki on August 18, 2020, 06:51:32 pm ---You’re not going to get a reply from an engineer. Expecting a reply like that is an exercise in frustration.

--- End quote ---


LOL, don't I know it! Well, when I review that product on Amazon and explain to people they are actually getting 70w atts RMS at 8 ohms, at least there's that. The seller actually admitted that in the public comments section, though.

I tried one of these dual 316D amps before and it seemed like it put out more power, not at a static 8 phms, but just turning up the dial listening to music, it seems to have a lot more punch. No way to tell.

The problem I had is that it was noisy when turning up the volume knob and with no music playing. It had a lot of hiss. My single amp 3116D unit does not have that. It's really quite. It was the same vendor's product.

This new unit I'll test as well as I can tonight for hiss or any other artifacts. It's also a Bluetooth 5.1 receiver and as stated before, has a passive sub out for an active sub connection. I think this is where any of these 3116D chips could really shine, when you use the 3116D power to power mids and highs only, and use an active sub for the rest

However, you could do the same thing with a single chip and add a passive sub out to it.

DW1961:
I had an epiphany!

Someone would have to do the calculation, maybe I will.

But! I think I have figured out one  benefit  from using (2) 3116D chips.

First, I just want to make sure my calcs are right about this included PSU. The included PSU for this unit is 24V/4.5A. That means the most power you will get with that PSU at 2 ohms is 40 watts total--lol? You need 10 A to get 200 watts at 2 ohms. Am I seeing that correctly?

Anyway, here's my idea. Let's keep the voltage at 24V even though the specs say you can use 26V, and let's allow amps to be whatever they are. (Did anyone actually see an amp rating for these chips? I'm assuming it's 7 A because it needs 7 amps at 2 ohms in mono to make 100 watts.)

You could wire (4) 4 ohms speakers in parallel and get 2 ohms or you could use (4) 8 ohms speakers at 4 ohms. Using a single chip, the most power you will get at 24V and 4 ohms is 100 watts.

If you wire up 4-8 ohms speakers you get get a 4 ohm speaker system at 100 watts shared, or 25 per channel.

FOUR SPEAKER SYSTEM
4 0hms
5 A
20 V
= 100 watts.
25 watts per speaker

TWO  SPEAKER SYSTEM
8 ohms
3A
24V
= 72 watts
36 watts per speaker.

You lose 30% of your power, but you gain 100% more speaker. Just off the top of my head, you're going to make a lot more volume with the 4 speaker setup at 4 ohms than using 2 at 8 ohms (depending on speaker sensitivity).

Am I far enough down the rabbit hole yet?

rsjsouza:

--- Quote from: cv007 on August 18, 2020, 06:56:29 am --->In this case, I would assume the relationship between figures 14 and 22 could also be applied to figure 13, that is: at 24V, power at 8 ohms in PBTL would be double when compared to BTL

You are missing the fact that the PBTL graph is using 2 Ohms. All the graphs essentially end up the same (BTL 8,4,PBTL 2), each is just changing the load impedance to increase the power. PBTL is already BTL, its just parallel to share the heat.

--- End quote ---
I am not missing anything. I am just using the output power characterized in Figure 13 (42-ish W for 24V in 8 ohms) and doubling it for PBTL, just like I did when extrapolating figure 14 to PBTL in 4 ohms (which coincidentally matches 2 ohms in PBTL).

Oh well... Take it as it is. I would go with the manufacturer's specs and characterization curves instead of assuming that all is equivalent and equal.

Good luck with your amp, DW1961. Peace out.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod