Led hacking help for robotank

[bigQQ]

That’s technically not what I asked, since I already know this ^^

I wanted to know how the mt7201 is driven in this specific lamp, because currently it looks like the circuit can’t dimm below 25%, which means either direct PWM above 20kHz or PWM smoothed to DC by the components next to the mt7201.

The PWM signal is a duty cycle. My control chip uses a 15k signal, which can stably light the LED. It is actually used. The output current is controlled by the signal. The voltage is determined by the load,
pwm信号是占空比的方式，我的控制芯片采用的15k的信号，可以稳定点亮led，是经过实际使用的，通过信号控制输出的电流大小，至于电压是根据负载决定的，

 

[bigQQ]

That’s technically not what I asked, since I already know this ^^

I wanted to know how the mt7201 is driven in this specific lamp, because currently it looks like the circuit can’t dimm below 25%, which means either direct PWM above 20kHz or PWM smoothed to DC by the components next to the mt7201.

PWM信号是占空比。我的控制芯片使用15k信号，可以稳定地点亮LED。它是实际使用的。输出电流由信号控制。电压由负载确定，The PWM signal is a duty cycle. My control chip uses a 15k signal, which can stably light the LED. It is actually used. The output current is controlled by the signal. The voltage is determined by the load,

 

[Cantankerous]

Thankyou so much for your advice everyone. Apologies for being AWOL, Christmas with a newborn was absolutely crazy and then got back into the office and my boss was making life a misery, they didn't hire anyone when I was away on paternity leave, just left work to pile up then second I stepped in the door on my case about the backlog . I'm hoping to have some time this weekend to read through everything and understand it properly. Thankyou again for this awesome advice, can't tell you how much it helps to have, particularly now I'm getting a bit time poor. Not having to research everything from scratch is a huge shortcut, massively appreciated.

@bigQQ have you modified some of these lights before? Amazed to find someone who has worked with them, in my area they are the secret weapon of the professional sellers but not widely known beyond them. I got very lucky sourcing mine. They were the display lights of several of the top fish shops in my area. They'd sell other brands but use these for their own tanks. When they changed their lights out I managed to get a hold of 2, then the owner of another store was selling his personal aquarium and I got 4 more.

Great colour spectrum, but very tricky to program. Chinese interface on the box, and a lot of issues connecting to networks. Similar issues to the jebao products, I think china may have different wifi protocols or something, but just really hard to connect to wifi. If they will connect, when they lose connection they then refuse to reconnect and have to start over each time. Having the hardware from these lights controlled by a western interface that connects reliably, like the reefpi, would be perfect.

 

[oreo54]

Thankyou so much for your advice everyone. Apologies for being AWOL, Christmas with a newborn was absolutely crazy and then got back into the office and my boss was making life a misery, they didn't hire anyone when I was away on paternity leave, just left work to pile up then second I stepped in the door on my case about the backlog . I'm hoping to have some time this weekend to read through everything and understand it properly. Thankyou again for this awesome advice, can't tell you how much it helps to have, particularly now I'm getting a bit time poor. Not having to research everything from scratch is a huge shortcut, massively appreciated.

@bigQQ have you modified some of these lights before? Amazed to find someone who has worked with them, in my area they are the secret weapon of the professional sellers but not widely known beyond them. I got very lucky sourcing mine. They were the display lights of several of the top fish shops in my area. They'd sell other brands but use these for their own tanks. When they changed their lights out I managed to get a hold of 2, then the owner of another store was selling his personal aquarium and I got 4 more.

Great colour spectrum, but very tricky to program. Chinese interface on the box, and a lot of issues connecting to networks. Similar issues to the jebao products, I think china may have different wifi protocols or something, but just really hard to connect to wifi. If they will connect, when they lose connection they then refuse to reconnect and have to start over each time. Having the hardware from these lights controlled by a western interface that connects reliably, like the reefpi, would be perfect.

Well welcome back..
As much as I love puzzles you may want to consider just replacing the entire "drivetrain"
Attach ps and led wires..along with a know easy to use dimming protocol.
$100-ish per light for 10 drivers, 2 boards and a fan module.
Done.
This isn't the cheapest place for parts either.
Board has spot for a $10 fan "driver" (i.e step down to 12v )
You can use ldd-l's (limited to a 36 or less voltage ps)
and a different board. There is someone w/ a surplus of driver boards somewhere here.

"Fitting" may be an issue and you will need to figure the current of each channel @ 100%.
Going a bit over or under shouldn't be an issue.
I still can't verify the useable pwm voltage but my guess and change from prev post it isn't 3.3 or 5V.

The "weirdness" in the dim response.. i.e changing current than the set point depending on "stuff" seems err..annoying.
I went to a few of my "go to" electronics sites and found really nothing on these driver ic's..
I did find them used in led flashlights and possibl automotive lighting. Neither of which required dimming.

Oh looks like there is room for some form of ldd be it on a board or discreet wired drivers (Ldd-hw or lw)

 

[oreo54]

Lingwendil is the member who has 4 up ldd- l and 8 up ldd- h boards btw

Those driver chips are like 17 cents each at Alibaba .

 

[bigQQ]

Thankyou so much for your advice everyone. Apologies for being AWOL, Christmas with a newborn was absolutely crazy and then got back into the office and my boss was making life a misery, they didn't hire anyone when I was away on paternity leave, just left work to pile up then second I stepped in the door on my case about the backlog . I'm hoping to have some time this weekend to read through everything and understand it properly. Thankyou again for this awesome advice, can't tell you how much it helps to have, particularly now I'm getting a bit time poor. Not having to research everything from scratch is a huge shortcut, massively appreciated.

@bigQQ have you modified some of these lights before? Amazed to find someone who has worked with them, in my area they are the secret weapon of the professional sellers but not widely known beyond them. I got very lucky sourcing mine. They were the display lights of several of the top fish shops in my area. They'd sell other brands but use these for their own tanks. When they changed their lights out I managed to get a hold of 2, then the owner of another store was selling his personal aquarium and I got 4 more.

Great colour spectrum, but very tricky to program. Chinese interface on the box, and a lot of issues connecting to networks. Similar issues to the jebao products, I think china may have different wifi protocols or something, but just really hard to connect to wifi. If they will connect, when they lose connection they then refuse to reconnect and have to start over each time. Having the hardware from these lights controlled by a western interface that connects reliably, like the reefpi, would be perfect.

My English is not good. I use Google Translator. I'm from a Chinese manufacturer. I made this lamp, so I know better. If you have any questions or suggestions, you can tell me that we have made improvements to the network problems you have raised. It is more convenient and fast to use a new Bluetooth connection. It is also very convenient to set up. If you need some lamp accessories, I can provide them to you.我的英文不好，我都是用谷歌翻译器，我是来自中国的厂家，这款灯是我做的所以我比较了解，如果你们有什么问题或者建议，都可以跟我说，针对你提出的网络问题，我们已经做了改善，采用新的蓝牙连接的方式，更加的方便快捷。设置起来也很方便。如果需要一些灯具的配件，我可以提供给你们。

 

[bigQQ]

Well welcome back..
As much as I love puzzles you may want to consider just replacing the entire "drivetrain"
Attach ps and led wires..along with a know easy to use dimming protocol.
$100-ish per light for 10 drivers, 2 boards and a fan module.
Done.
This isn't the cheapest place for parts either.
Board has spot for a $10 fan "driver" (i.e step down to 12v )
You can use ldd-l's (limited to a 36 or less voltage ps)
and a different board. There is someone w/ a surplus of driver boards somewhere here.

"Fitting" may be an issue and you will need to figure the current of each channel @ 100%.
Going a bit over or under shouldn't be an issue.
I still can't verify the useable pwm voltage but my guess and change from prev post it isn't 3.3 or 5V.

The "weirdness" in the dim response.. i.e changing current than the set point depending on "stuff" seems err..annoying.
I went to a few of my "go to" electronics sites and found really nothing on these driver ic's..
I did find them used in led flashlights and possibl automotive lighting. Neither of which required dimming.

Oh looks like there is room for some form of ldd be it on a board or discreet wired drivers (Ldd-hw or lw)

In view of the price problem you have raised, all the circuit boards of this R80 are priced at 288 yuan in China, and the light boards are priced at 678. We have designed a board that considers many things. The fan module, temperature module and time module are integrated together, and can be linked through Bluetooth and wifi. There are plenty of materials. In fact, there are a lot of cheap circuit boards. They don't have chips, but they are running programs in a virtual way. We have changed to a more functional chip scheme. At the price of $100, we can make a smaller lamp, It will cost 1100 yuan to make a set of lamps like R80 in China. Now the price of R80 finished lamps in China is 1800 yuan. There is also a new X240, which is closer to the xr30 g5 blue lamp. The price is 3000 RMB. If you like Diy, I can help you find any accessories you need, and I can also find those with lower prices.针对大家提出的价格问题，这款R80的所有电路板在国内售价为288元，灯板售价为678元。我们设计了一个考虑很多事情的电路板。风扇模块、温度模块和时间模块集成在一起，可以通过蓝牙和wifi进行链接。有很多材料。事实上， 有很多便宜的电路板.他们没有芯片，但他们以虚拟方式运行程序。我们已经改为功能更强大的芯片方案。以100美元的价格，我们可以做一个更小的灯，在中国做一套像R80这样的灯需要1100元。现在R80成品灯在中国的价格是1800元。还有一款新的X240，更接近XR30 g5蓝灯。价格为3000元。如果你喜欢DIY，我可以帮你找到你需要的任何配件，我也可以找到价格更低的配件。

 

[Sral]

Well welcome back..
As much as I love puzzles you may want to consider just replacing the entire "drivetrain"
Attach ps and led wires..along with a know easy to use dimming protocol.
$100-ish per light for 10 drivers, 2 boards and a fan module.
Done.
This isn't the cheapest place for parts either.
Board has spot for a $10 fan "driver" (i.e step down to 12v )
You can use ldd-l's (limited to a 36 or less voltage ps)
and a different board. There is someone w/ a surplus of driver boards somewhere here.

"Fitting" may be an issue and you will need to figure the current of each channel @ 100%.
Going a bit over or under shouldn't be an issue.
I still can't verify the useable pwm voltage but my guess and change from prev post it isn't 3.3 or 5V.

The "weirdness" in the dim response.. i.e changing current than the set point depending on "stuff" seems err..annoying.
I went to a few of my "go to" electronics sites and found really nothing on these driver ic's..
I did find them used in led flashlights and possibl automotive lighting. Neither of which required dimming.

Oh looks like there is room for some form of ldd be it on a board or discreet wired drivers (Ldd-hw or lw)

True, just using pre-made drivers should be a breeze, except on money maybe :grinning-face-with-sweat:

Current per channel is easy though, since maximum current per driver is 1.0 A @RSENSE of 0.1 Ohm (R100 on the board). So as you mentioned previously we can be pretty certain that the drivers are driven as per datasheet at:

SMD resistor	Resistance	max Current
R100	0.10 Ohm	~1000 mA
R130	0.13 Ohm	~ 770 mA
R300	0.30 Ohm	~ 333 mA

I absolutely agree with you on the PWM. Driving the circuit with anything but 500 Hz PWM seems absolutely unreasonable, since PWM is easy for microcontrollers and PWM gives the best dim response.

 

[Sral]

My English is not good. I use Google Translator. I'm from a Chinese manufacturer. I made this lamp, so I know better. If you have any questions or suggestions, you can tell me that we have made improvements to the network problems you have raised. It is more convenient and fast to use a new Bluetooth connection. It is also very convenient to set up. If you need some lamp accessories, I can provide them to you.我的英文不好，我都是用谷歌翻译器，我是来自中国的厂家，这款灯是我做的所以我比较了解，如果你们有什么问题或者建议，都可以跟我说，针对你提出的网络问题，我们已经做了改善，采用新的蓝牙连接的方式，更加的方便快捷。设置起来也很方便。如果需要一些灯具的配件，我可以提供给你们。

Thanks, good to know !

I think the biggest problem with the lamp for our use case is, that we are missing an easy to use interface with third-party control devices, like ReefPi. WiFi and Bluetooth is great, but it always limits the use to using the manufacturers APP, which typically doesn't disclose the WiFi or Bluetooth protocol so one could control the lamp directly with third-party control devices like ReefPi. Alternatively, a PWM input for external control signals would also be great, although I'm afraid that this is unnecessary for almost all users and would therefore unnecessarily increase cost.

It would be absolutely great if you could provide us with circuit diagrams of the driver board, or at least with a rough sketch. Alternatively the WiFi protocol of the control board would also be usefull, so we could think about controlling the lamp over WiFi directly from ReefPi.

 

[bigQQ]

Thanks, good to know !

I think the biggest problem with the lamp for our use case is, that we are missing an easy to use interface with third-party control devices, like ReefPi. WiFi and Bluetooth is great, but it always limits the use to using the manufacturers APP, which typically doesn't disclose the WiFi or Bluetooth protocol so one could control the lamp directly with third-party control devices like ReefPi. Alternatively, a PWM input for external control signals would also be great, although I'm afraid that this is unnecessary for almost all users and would therefore unnecessarily increase cost.

It would be absolutely great if you could provide us with circuit diagrams of the driver board, or at least with a rough sketch. Alternatively the WiFi protocol of the control board would also be usefull, so we could think about controlling the lamp over WiFi directly from ReefPi.

Sorry, the agreement of wifi cannot be provided to you. It doesn't help if you know. His protocol is made with my app, so if you want to provide pwm signals to the driver board, you can use the reefpi to provide them, and you can control the frequency.很抱歉，wifi的协议是不能提供给你的。你知道了也没有用，他的协议是配合我的app做的，所以你想给驱动板提供pwm信号，可以用你说的reefpi去提供，频率可以你来掌控。

 

[Sral]

Sorry, the agreement of wifi cannot be provided to you. It doesn't help if you know. His protocol is made with my app, so if you want to provide pwm signals to the driver board, you can use the reefpi to provide them, and you can control the frequency.很抱歉，wifi的协议是不能提供给你的。你知道了也没有用，他的协议是配合我的app做的，所以你想给驱动板提供pwm信号，可以用你说的reefpi去提供，频率可以你来掌控。

Sad to hear, but I was afraid of that.

Can you at least tell us the wiring of the components between the WiFi control chip's PWM pin and the mt7201's ADJ-pin ?
If you can't disclose the diagram, can you at least tell us whether there are any series resistors in the line or capacitors parallel to it like this:

 

[bigQQ]

Sad to hear, but I was afraid of that.

Can you at least tell us the wiring of the components between the WiFi control chip's PWM pin and the mt7201's ADJ-pin ?
If you can't disclose the diagram, can you at least tell us whether there are any series resistors in the line or capacitors parallel to it like this:

Well, there is no original. The signal output by the chip is directly connected to the drive port,

 

[Sral]

Well, there is no original. The signal output by the chip is directly connected to the drive port,

Thanks a lot !
Just to be sure, since it could lead to damaging results when we replicate this incorrectly:
does your control chip output a 3.3V PWM with 0-100% duty cycle or does it have an open drain and the voltage is defined by the mt7201 ?

 

[bigQQ]

Thanks a lot !
Just to be sure, since it could lead to damaging results when we replicate this incorrectly:
does your control chip output a 3.3V PWM with 0-100% duty cycle or does it have an open drain and the voltage is defined by the mt7201 ?

Yes, 0-100, 3.3v

 

[oreo54]

Yes, 0-100, 3.3v

Did you measure the output current at 100% to verify it matches the expected setpoint?

Sorry just a curiosity on my part.

 

[Sral]

Did you measure the output current at 100% to verify it matches the expected setpoint?

Sorry just a curiosity on my part.

Exactly … the mt7201 ADJ port has a max voltage of 1.25V for 100%current, if you give it 3.3V at 100% it gets cropped to 2.5V by the internal diodes and delivers 200%, which is 2.0A and way too much.

although it would line up with @Cantankerous‘ previous measurements. Strange, somebody isn’t telling the truth here, either our informant or the datasheet :grinning-face-with-sweat:

 

[oreo54]

Exactly … the mt7201 ADJ port has a max voltage of 1.25V for 100%current, if you give it 3.3V at 100% it gets cropped to 2.5V by the internal diodes and delivers 200%, which is 2.0A and way too much.

Funny you should post that at this time.
I was just looking over the spec sheet for my personal "electronics 101" session.
:)
Anyways what I "see" is the chip has an internal reference signal of 1.2V. I assume that is to keep the chip "on"
if no input on the ADJ pin.
Second it contains a lowpass filter (R/C) to convert a pwm to a dc voltage (triangle wave??)

not sure about where the diodes kick in here anyways, assuming a full 2.4V dc applied to the pin the equation is

The nominal average output current in this
case is given by:
I OUTdc = V ADJ /1.2/RS [for 0.3<V ADJ <2.5V]
Note that 100% brightness setting
corresponds to VADJ =V REF.

2.4/1.2/.1 (for 1000ma) Which is 20??? when it should be 2
1.2/1.2/.1 1/.1??
Have I lost all my math skills?

current IOUTnom=0.1/RS

Anyway guessing .. at 3.3V DC on the adj pin and a .1 Ohm sense resistor (ignoring pwm atm) the output current would be 2.75A :eek:
except for the diodes? Exactly where are they?
And if it did this (2.75A) "in practice" I assume either the driver or led string would not last long.

I've dabbled in using the voltage drop in diodes to cut the voltage down but that is external.

I guess one could compensate by using different sense resistors..
Say you wanted 1A max when applying 3.3v dc to adj pin one would use a .275 Ohm sense resistor.
.1/.275 = .363A x 2.75 =1A ?

I'm missing something.. AFAICT the max voltage on the Adj pin should be 1.2
3.3v regardless of pwm or not just seems wrong. Even averaging the voltage over the duty cycle
still puts higher than spec-d voltage especially at 100%.

Oh also going to reiterate that unless there is an overriding reason the PWM frequency should be <500Hz
A 100:1 brightness control range sure beats 5:1
Cameras may not like it as much though.

since this driver seems to actually adj. current (unlike say an LDD which has a PWM output so current is equal at all on times) there will be some color shift as you dim.

 

[Sral]

Funny you should post that at this time.
I was just looking over the spec sheet for my personal "electronics 101" session.
:)
Anyways what I "see" is the chip has an internal reference signal of 1.2V. I assume that is to keep the chip "on"
if no input on the ADJ pin.
Second it contains a lowpass filter (R/C) to convert a pwm to a dc voltage (triangle wave??)

not sure about where the diodes kick in here anyways, assuming a full 2.4V dc applied to the pin the equation is


2.4/1.2/.1 (for 1000ma) Which is 20??? when it should be 2
1.2/1.2/.1 1/.1??
Have I lost all my math skills?



Anyway guessing .. at 3.3V DC on the adj pin and a .1 Ohm sense resistor (ignoring pwm atm) the output current would be 2.75A :eek:
except for the diodes? Exactly where are they?
And if it did this (2.75A) "in practice" I assume either the driver or led string would not last long.

I've dabbled in using the voltage drop in diodes to cut the voltage down but that is external.

I guess one could compensate by using different sense resistors..
Say you wanted 1A max when applying 3.3v dc to adj pin one would use a .275 Ohm sense resistor.
.1/.275 = .363A x 2.75 =1A ?

I'm missing something.. AFAICT the max voltage on the Adj pin should be 1.2
3.3v regardless of pwm or not just seems wrong. Even averaging the voltage over the duty cycle
still puts higher than spec-d voltage especially at 100%.

Oh also going to reiterate that unless there is an overriding reason the PWM frequency should be <500Hz
A 100:1 brightness control range sure beats 5:1
Cameras may not like it as much though.

since this driver seems to actually adj. current (unlike say an LDD which has a PWM output so current is equal at all on times) there will be some color shift as you dim.

Maybe we have been consulting the wrong datasheet. This one for the MT7201C+ mentions a higherDC ADJ voltage that is also clamped to a maximum current of 100%. It also says that the nominal PWM amplitude is supposed to be between 2.5 and 5.5 V.

If that’s the circuit our informant was absolutely right. One way To check would be to power the driver board and measure the voltage on the disconnected PWM_IN. If that’s 2.4V we have the MT7201C circuit with a different ADJ-behavior :grinning-face-with-sweat:

And can therefore simply drive it with a 3.3V PWM @ 500 Hz. Case closed XD

 

[bigQQ]

Did you measure the output current at 100% to verify it matches the expected setpoint?

Sorry just a curiosity on my part.

Yes, after verification, this scheme has been used by R80 coral lamp

 

[bigQQ]

Exactly … the mt7201 ADJ port has a max voltage of 1.25V for 100%current, if you give it 3.3V at 100% it gets cropped to 2.5V by the internal diodes and delivers 200%, which is 2.0A and way too much.

although it would line up with @Cantankerous‘ previous measurements. Strange, somebody isn’t telling the truth here, either our informant or the datasheet :grinning-face-with-sweat:

I don't know how you measured the data. I can say for sure that the pwm signal is 0-100 and the voltage is 3.3v. Then the 7201 chip supports a maximum current of 1a. The final output voltage of the port is determined by the load lamp bead,我不太了解你们怎么测量出来的数据，我可以很肯定的说，pwm信号是0-100，电压3.3v。然后7201这个芯片最大支持1a电流。端口的最终输出电压是根据负载灯珠决定的，