Using SMPS near DWM1001 Module

Hi,

I am trying to run DWM1001 with an SMPS of 5V 1A output. I am currently using IRM-05-5 from MeanWell.https://www.mouser.in/datasheet/2/260/r1936_3-1109422.pdf

What about the effect of EMI if I place the SMPS near DWM1001. What is the safe distance between SMPS and DWM. What are the PCB Design rules if I would like to place SMPS near DWM1001.

Any help around the EMI/EMC effects due to power electronics would be great .

Thanks,
Nav

Hi Nav,
You can’t power the DWM1001 from 5 V as it requires a 3.3 V supply. Also, the DW1000 IC can be sensitive to power transients when entering transmit and receive modes, so you should select a DCDC with fast transient response. The part you’re using from MeanWell seems to be quite a heavy duty part, almost like an AC adapter and I’m not sure how suitable it is. But as it outputs 5 V, you’re going to need a regulator to take that down to 3.3 V anyway. You could use Torex XC9258B33, which we have validated for use with DWM1001. It’s listed on the DW1000 datasheet as a recommended part also.
For layout guidelines regarding EMI, it’s always important to follow the vendors specific guidelines. Usually you need to ensure power inductor and capacitor components are kept as close as possible to the regulator IC, use wide PCB traces, etc. to ensure current return loops are minimised.

1 Like

Hi S_Carrol,

Thanks for the reply.

I am using DC/DC between AC SMPS and DWM1001. Currently I am using RT8509. I am keeping capacitors and inductors close to regulator.
I have provided sufficient space for Antenna and created a safe isolation area around Antenna from ground/Metal.
If I keep SMPS and the DWM1001 on the Top layer, 30mm apart will have any effect on radio?
any suggestions around this would be great help.

Thanks,
Nav

Hi Nav,

Actually, when we evaluated the Richtek RT8509, the transient response wasn’t fast enough and we occasionally saw packet errors in our UWB link. So I would recommend you that you change to the Torex XC9258, which we have seen shows no issues.

Once your PCB layout of the DCDC is correct, then you don’t really need to worry too much about the proximity. It makes sense not to place it close to the antenna. 30 mm clearance is certainly plenty.

Thanks S_Carrol. Appreciate your help.

Regards,
Nav

Hi,
I am trying to use Torex XC9258B33 for 3.3V supply.In that for selecting inductor and capacitor values two fosc conditions are mentioned.Which values can we use for getting 3.3V?Which oscillation frequency condition can we use?

Inductor: https://www.digikey.in/product-detail/en/taiyo-yuden/NR3015T4R7M/587-1649-1-ND/1008264.
Capacitor: 10uF, 10V
I have used this. It works well.

Nav

ok Thanks for the reply

XC9258 has two switching frequency variants. The slower switching speed (1.2 MHz) has slightly better power efficiency, while the faster switching speed (6 MHz) would have slightly faster transient response. So there’s a trade-off when it comes to selecting DCDC converter switching speed.

Dear Nav,

For the PCB board design, is two layer will be enough or do we need to design by Four Layer PCB board.

Thanks & Regards
Suman

While a 2 layer DW1000 design is theoretically possible, my expectation is that it will be fraught with problems and is highly unlikely to function adequately. Even 4 layer designs can have issues over what seem like trivial variations, such as number and pattern of vias to the ground plane under the DW1000 chip.

The cost difference between 2 and 4 layer is not that great so make your design 4 layers. The cost savings for 2 layers will be eaten up by the extra design time and debugging frustration at the end, and you are going to end up building again in 4 layers anyway. In volume production, the cost difference is rather small, well under $1 per unit.

Ideally, you should look at an existing known working design where you have the PCB file or the Gerbers and follow the layout example closely. Basically, it is all pretty critical at 6.5 GHz. Pay attention to the layer stack up as well so the traces have the right impedance. Also, board material makes a measurable difference, don’t just specify generic FR-4, you need to be specific as to grade.

It is schedule and budget busting to tell a client that their in-house board design is not salvageable and it has to be redone. We’ve had to do that a number of times over the last 6 years we have worked with Decawave. 6+ GHz design is not for the faint of heart. The electrons cannot be reasoned with and will embarrass you at every opportunity, so you have to be very diligent to keep them in their place.

Mike Ciholas, President, Ciholas, Inc
3700 Bell Road, Newburgh, IN 47630 USA
[email protected]
+1 812 962 9408

Hi Suman,

If your using DWM1001 then 2 layer works well.

Thanks,
Nav

Hi,
How to integrate battery status circuit for the DWM1001 module .we are not using any battery charging circuit.

use A2D or Comparator circuit to sense the percentage of batteries.

Nav

Hi S_Carrol,

Can the 3.3V and 1.8V power supply of DW1000 IC use LDO? For example, Torex XC6228, Richtek RT9193, TI TLV70233.

LDOs can work. LDOs are going to be less noise and complexity than switching regulators.

But, they come with some caveats. The obvious one is the LDOs waste more energy so if you are battery powered, you need a bigger battery for the same run time, or you have less run time for the same size battery. If the battery is, say, 3.7 volts for LiIon, then LDO to 1.8 volts is under 50% efficient. A good switching regulator can get that to above 90% for comparison.

A related issue is that the LDO will generate heat. For a DW1000 in receive, about 400 mW, the LDO is generating about 400 mW of heat as well, and if the LDO is small, the temperature can get quite high. In some cases, the thermal treatment to make an LDO work end up taking more board space than a switching regulator would have.

The heat can also cause weird side effects particularly for the node crystal. If a node is in receive a long time, the board temperature rises and that causes the crystal to change frequency. If the node is then put in a lower power mode, the temperature will change again and affect the crystal. So you can into situations where the crystal wandering around affects your accuracy depending on your use pattern. For the best anchor performance, we keep the DW1000 generally in receive so it reaches a stable temperature. When you design a battery powered anchor, this is not possible, so temperature effects become more prominent.

Lastly, LDOs, particular the ones who claim very low quiescent currents, often have poor transient response. That is, they don’t respond well to sudden load changes. The DW1000 causes sudden load changes when it changes mode, and even each transmit pulse is a short term load change. This can mean certain LDOs are simply not usable for a DW1000 system. So don’t get excited about a low Iq until you check out the transient response curves.

To be clear, I have not evaluated the particular LDOs you nominated.

Mike Ciholas, President, Ciholas, Inc
3700 Bell Road, Newburgh, IN 47630 USA
[email protected]
+1 812 962 9408

Hello,S_Carroll,

RT8509 is a boost DCDC, which is not suitable for use here. I think you are talking about RT8059. Then I compared the transient response of RT8059 and XC9258, and I found that the parameters of RT8509 are better than XC9258.
Now I want to choose a DCDC IC with a 4.2V lithium battery to convert to 3.3V to power the DWM1000. I have considered RT8059, XC9058 and MP2162, but I don’t know which one to choose.

Hi harry_z,

Sorry, RT8059 is the correct part number. Excuse my typo in my comment above.

When it comes to transient response of a DCDC supplying the 3V3 of DW1000, it’s very subtle. Yes, the RT8059 performance looks comparable to XC9258 but we see that the load current is different, which makes the comparison unfair (often, a regulator can turn on and stabilise quickly in PWM mode when the load current is nice and high). And of course the DW1000 turning on to transmit or receive is a unique and dynamic load, as opposed to what a DCDC vendor has tested with (a constant resistive load, most likely). So you really have to do your own evaluation if you want to use a new DCDC on the 3V3 rail, with the DW1000 waking up from sleep state to transmit or receive, as it will in your end application. And you need to check that the key performance indicators such as receiver sensitivity and ranging accuracy (standard deviation) are good for your chosen power solution. We’ve gone through this process with RT8059 and XC9258, so you could save alot of time by going with our recommendation to use XC9258.

But you should also take a step back and consider if a DCDC is really needed. Your lithium battery will spend most of the time at 3.7 V. Using a 3.3 V LDO would give you around 89% efficiency at this voltage (okay, as the battery decays below 3.3 V, you will lose the noise rejection capability of the LDO but if there aren’t any major noise sources nearby, that should be okay).

If using a DCDC, you just need to be aware that, whilst the efficiency is good when the DW1000 is transmitting or receiving (high load current), it will be poor when the current is very low (in sleep states).

Consider how much time the device will spend in active (Tx and Rx) vs sleep states each day, do some battery lifetime calculations and let this guide your decision.

Hi S_Carroll_DW.
What about the CE pin of XC9258.
The datasheet refers to a resistor not more than 1M between CE and VIN or VSS, to put into “H” or “L”.
Any recommendation?

Thanks