Output not stable

Tekemachi

on Jun 7, 2024

Category: Hardware

Product Number: LTC7804

Hello,

I am designing a boost converter +12V to +16V/10A.

I have a sawtooth waveform at the output at 0 load (see picture below):

And at high load, the output capacitor gets really hot.

I have also try adding 10R gate resistors to both MOSFET but no luck.

This is my schematic: PDF

Can you help me please.

Thank you in advance.

Tags: power management hardware LTC7804 Step-Up (Boost) Regulators sawtooth Micropower Boost Regulators Show More

0 Achim
on Jun 8, 2024 3:38 AM
Hello Tekemachi,

Tekemachi said:
I have a sawtooth waveform at the output at 0 load (see picture below):

You have connected the mode pin for burst mode operation at low load. In that case it's normal to get an increased voltage ripple when there is no load at the output.

Is this increased volatge ripple really be a problem for you? (In most cases it's not..). If it is a problem, you could instead use the forced continuous mode. That will reduce the voltage ripple at no load, but it will give you a higher current consumption at no load/low load.

Tekemachi said:
And at high load, the output capacitor gets really hot.

I did not check the calculations of your tool for L and Cout. Comparing it to the datasheets example circuits both values seem quite small for, but as I said: I did not calculate myself. But anyhow: this is a boost converter, and for boost converters the ouput capacitor has to withstand a high AC current. Your tool says, that 11 Arms flow through Cout, so you need capacitors that can withstand 11 A at 1 MHz.

Do you have ESR-values for your capicotors at 1 MHz? Which specific components did you choose for C144 and for C145-C147? (i.e. not only the capacitance value and voltage rating, but the specific part number). It's obviously not the 5x47µF TDK-capacitors that your design tool proposed.

best regards

Achim
Cancel
Up 0 Down

Reply

Verify Answer

Cancel
0 Tekemachi
on Jun 9, 2024 7:15 AM in reply to Achim
Hello Achim,

Thank you for your reply.

I'm using the boosted voltage to supply a motor controller DRV8873H from TI. The input waveform to control the motor is 200Hz. I will try to use the forced continuous mode and see.

Further than that, do you see and mistakes in the design?

These are the part values:

C144: A768KE226M1JLAE054

C145-C147: CL32B106KMVNNWE

C144 gets really hot even and light load about 2A.

If I remove C144, the boost converter doesn't work at all.

I will also try to get the recommended capacitor and see.

Thank you for your help.
Cancel
Up 0 Down

Reply

Verify Answer

Cancel
0 Achim
on Jun 10, 2024 3:22 PM in reply to Tekemachi
Hello Tekemachi,

Tekemachi said:
These are the part values:

C144: A768KE226M1JLAE054

C145-C147: CL32B106KMVNNWE

C144 gets really hot even and light load about 2A.

C144 has a specification for ESR and for allowed ripple current up to 1 MHz. ESR is said to be 54mOhm. That's not bad, but if the complete ripple current of your converter would flow through this cap, it would heat up with

P = (11A)^2 * 54mOhm = 6,5W.

That would for sure be much too much. Even a ripple current of just 2 A might be too much: the specification of C144 allows a ripple current of 1,175A when the cap is hot.

So we have to hope, that C145-C147 carry most of the ripple current. Here you can see detailed specification for these caps.

https://product.samsungsem.com/mlcc/CL32B106KBJNNN.do

The combined capacity of C145-C147 is nominally 30µF, but it is alread reduced to 22,5µF with an output voltage of 16V. If we just take these capacitance numbers, 50% of the ripple current would flow through C144, and this will be destructive for C144 at high output currents.

Tekemachi said:
Further than that, do you see and mistakes in the design?

As I said before: I'm not sure on what basis your tool calculated the required values for L and for Cout. Compared to the typical circuits in the datasheets, the values in your schematic seem quite small to me.

Besides that: for a 1 MHz switching converter with currents up to 18 A, not only the schematic has to be correct, the layout also needs to be well thought out. If you should have unfavorable current loops in the region of M3, M4 and C144-C147, the behavior of Cout can be influenced by unwanted parasitic inductivities. And Cout is one of the most critical components for such a high current boost converter.

Tekemachi said:
I'm using the boosted voltage to supply a motor controller DRV8873H from TI.

I cannot really recognise the absolute value of the current ripple in your scopeshot (the picture made with the mobile phone in portrait orientation cuts off the scaling information, which is probably shown somewhere on the screen.) But there is one measurement result saying peak-peak value (of voltage ripple?) is 720mV. Your DRV8873H works from 4,5V to 40V. I think, the DRV8873H won't care about 720mV ripple on a 16V supply.

Tekemachi said:
C144 gets really hot even and light load about 2A.

One further question from my side: where did you measure these 2A? Really at the ouput of the boost converter? Or on some other net? (The current on Vin is e.g. related to the current on Vout, but the absolute numbers differ).

best regards

Achim
Cancel
Up 0 Down

Reply

Verify Answer

Cancel

0 Tekemachi

on Jun 11, 2024 10:07 AM in reply to Tekemachi

Hello Achim , thank you so much for your answer.

I have attached my project here for your reference, I tried to minimize current loop with M1 and M2. The schematic for the boost converter is in sheet [2]-Power. Please ignore the sheet [3]-Peltier-Driver.

I am using the LTpowerCAD II v2 to calculate the values.

I am using a multimeter at the output for monitoring the current. I'm just using the circuit at 200Hz, so I think the multimeter is ok for this purpose. (don't be fooled by the project name, I am not driving a Peltier at 200Hz)

What I have tried:

I forced the circuit to operate at continuous mode by connecting MODE to INTVcc, it seems stables but rings horribly when I connect the load (about 1A) in: this is the output of the boost converter (16V)
I have tried removing C144 and put 3 more C145 but still, the circuit rings at about 1A of load current.

And you are correct, the ripple current is about 600mV.

I have attached my project and the LTpowerCAD (I changed it to .txt because the forum does not allow *.ltpc) file for your reference.

Here's also a screenshot:

Fullscreen 16V18A Boost - Copy.txt Download

<?xml version="1.0" encoding="utf-8"?>
<Root>
  <ConverterSpecs>
    <ConverterSpec>
      <Topology>Boost</Topology>
      <Type>Controller</Type>
      <OutputNum>One</OutputNum>
      <FswPreferred>1000</FswPreferred>
      <PartName>LTC7804</PartName>
      <IcTa>25</IcTa>
      <LastSavedFilesCultureInfo>de-DE</LastSavedFilesCultureInfo>
      <SolutionAreamm2>376,3</SolutionAreamm2>
      <SolutionComponentCount>11</SolutionComponentCount>
      <SolutionMaxHeightmm>8,9</SolutionMaxHeightmm>
      <TotalSolutionAreamm2>376,3</TotalSolutionAreamm2>
      <TotalSolutionPrice>7,7441</TotalSolutionPrice>
    </ConverterSpec>
  </ConverterSpecs>
  <ProgramUserInfos>
    <ProgramUserInfo>
      <LastRecordedSyncDate>10.04.2024 00:00:00</LastRecordedSyncDate>
      <ReleaseType>Customer</ReleaseType>
      <InstallationDate>03.08.2023 00:00:00</InstallationDate>
      <LTpowerCADVer>2</LTpowerCADVer>
      <LTpowerCADSubDesc>Release2</LTpowerCADSubDesc>
      <LTpowerCADUpdateVer>70</LTpowerCADUpdateVer>
      <LTpowerCADUpdateVerDate>27.04.2020 00:00:00</LTpowerCADUpdateVerDate>
    </ProgramUserInfo>
  </ProgramUserInfos>
  <OutputRailDatas>
    <OutputRailData>
      <PartTopology>Boost</PartTopology>
      <PartType>Controller</PartType>
      <PartLength>3</PartLength>
      <PartWidth>3</PartWidth>
      <PartHeight>0,8</PartHeight>
      <VinToCalc>12</VinToCalc>
      <IoToCalc>18</IoToCalc>
      <MaxVin>12</MaxVin>
      <NomVin>12</NomVin>
      <MinVin>12</MinVin>
      <SwitchingFreq>925000</SwitchingFreq>
      <OutputVol>16,08</OutputVol>
      <OutputCur>18</OutputCur>
      <NPh>1</NPh>
      <NIc>1</NIc>
      <VinD>12</VinD>
      <Ta>25</Ta>
      <InductorThetaWA>80,4</InductorThetaWA>
      <InductorRipplePercentDesired>50</InductorRipplePercentDesired>
      <Lo>2,5E-07</Lo>
      <DCR>0,00032</DCR>
      <InductorManufacturer>WURTH</InductorManufacturer>
      <InductorPartNumber>744301025</InductorPartNumber>
      <InductorLength>11,3</InductorLength>
      <InductorWidth>11</InductorWidth>
      <InductorHeight>8,9</InductorHeight>
      <InductorIsat>65</InductorIsat>
      <InductorIheat>40</InductorIheat>
      <InductorTheat>50</InductorTheat>
      <CinCeramic>1,0781777E-05</CinCeramic>
      <CinCeramicESR>0,00220000010449439</CinCeramicESR>
      <CinCeramicESL>4,80000068396333E-10</CinCeramicESL>
      <NCinCeramic>1</NCinCeramic>
      <CinCeramicManufacturer>TDK</CinCeramicManufacturer>
      <CinCeramicPartNumber>C2012X5R1V226M125AC</CinCeramicPartNumber>
      <CinCeramicPackage>0805</CinCeramicPackage>
      <CinCeramicLength>2</CinCeramicLength>
      <CinCeramicWidth>1,25</CinCeramicWidth>
      <CinCeramicHeight>1,25</CinCeramicHeight>
      <CinCeramicAtHalfRated>1,64100015588664E-06</CinCeramicAtHalfRated>
      <CinCeramicAtFullRated>5,08000048257412E-07</CinCeramicAtFullRated>
      <CinCeramicAtNoVoltage>2,20000020898879E-05</CinCeramicAtNoVoltage>
      <CinCeramicDerated>1,07817765344178E-05</CinCeramicDerated>
      <CinCeramicVoltageRating>35</CinCeramicVoltageRating>
      <CincPartPrice>0,4526</CincPartPrice>
      <VoRipplePkPkMaxTargetPercent>1</VoRipplePkPkMaxTargetPercent>
      <VoStepMaxTargetPercent>3,80382202325118</VoStepMaxTargetPercent>
      <TotalVoRippleMaxTargetPercent>4</TotalVoRippleMaxTargetPercent>
      <Coc>4,188541E-06</Coc>
      <ESRc>0,00156000007409602</ESRc>
      <ESLc>1,00000014249236E-11</ESLc>
      <Numc>5</Numc>
      <CeramicCapacitorManufacturer>TDK</CeramicCapacitorManufacturer>
      <CeramicCapacitorPartNumber>C3216X5R1E476M160AC</CeramicCapacitorPartNumber>
      <CeramicCapacitorLength>3,2</CeramicCapacitorLength>
      <CeramicCapacitorWidth>1,6</CeramicCapacitorWidth>
      <CeramicCapacitorHeight>1,6</CeramicCapacitorHeight>
      <CeramicCapacitorPackage>1206</CeramicCapacitorPackage>
      <CocAtHalfRated>5,15600048979373E-06</CocAtHalfRated>
      <CocAtFullRated>1,77800016890094E-06</CocAtFullRated>
      <CocAtNoVoltage>4,70000044647605E-05</CocAtNoVoltage>
      <CocDerated>4,18854119789004E-06</CocDerated>
      <CocVoltageRating>25</CocVoltageRating>
      <CocPartPrice>0,4293</CocPartPrice>
      <OutputLCFilterEnabled>False</OutputLCFilterEnabled>
      <OutputLCFilterSensingIsLocal>False</OutputLCFilterSensingIsLocal>
      <OutputLCFilterSensingIsRemote>False</OutputLCFilterSensingIsRemote>
      <Cth>4,70000066971409E-09</Cth>
      <Cthp>2,70000051297251E-11</Cthp>
      <Rth>2490</Rth>
      <RfbTopActual>60400</RfbTopActual>
      <RfbBotActual>4870</RfbBotActual>
      <FcDesired>65984,6</FcDesired>
      <PhaseBoostDesired>10</PhaseBoostDesired>
      <PhaseMarginDesired>60</PhaseMarginDesired>
      <Ilim>0,05</Ilim>
      <HotSwapCurrentLimitThresholdFoldbackFactor>1</HotSwapCurrentLimitThresholdFoldbackFactor>
      <MarginIoLimit>120</MarginIoLimit>
      <RsenseManufacturer>KOA SPEER</RsenseManufacturer>
      <RsensePartNumber>TLR3AWDTE1L50F75</RsensePartNumber>
      <Rsense>0,0015</Rsense>
      <RsenseTolerance>1</RsenseTolerance>
      <RsensePowerRating>2</RsensePowerRating>
      <RsenseType>METAL PLATE</RsenseType>
      <RsensePackage>2512</RsensePackage>
      <RsenseLength>6,35</RsenseLength>
      <RsenseWidth>3,18</RsenseWidth>
      <RsenseHeight>0,6</RsenseHeight>
      <RsenseFilterC>1E-09</RsenseFilterC>
      <SenseC>4,7E-08</SenseC>
      <SenseRp>7150</SenseRp>
      <SenseRs>4020</SenseRs>
      <CurrentSensingConfigurationType>1</CurrentSensingConfigurationType>
      <SwitchingFreqDesired>925000</SwitchingFreqDesired>
      <Rfreq>0,04</Rfreq>
      <HasExtvcc>True</HasExtvcc>
      <LoadStepHigh>4,5</LoadStepHigh>
      <LoadStepSlewRate>9999999</LoadStepSlewRate>
      <TonMin>8E-08</TonMin>
      <ProjectName>AIROBOX Peltier Power </ProjectName>
      <ProjectDate>10.04.2024</ProjectDate>
      <VinStartUp>12</VinStartUp>
      <VSD_Desired>12</VSD_Desired>
      <VOV_Latch_Desired>12</VOV_Latch_Desired>
      <VOV_NonLatch_Desired>12</VOV_NonLatch_Desired>
      <RailNumber>1</RailNumber>
      <BW>70,79</BW>
      <PM>60,34</PM>
      <DeltaVoLoadStep>433,654672582625</DeltaVoLoadStep>
      <TotalDeltaVoLoadStepPercent>2,69685741655861</TotalDeltaVoLoadStepPercent>
      <FootprintClearanceFactor>1,5</FootprintClearanceFactor>
      <Mode>Boost</Mode>
      <ModeChanged>False</ModeChanged>
      <PartName>LTC7804</PartName>
      <CapacitorDeratingEnabled>True</CapacitorDeratingEnabled>
      <RdsOnDeratingEnabled>False</RdsOnDeratingEnabled>
      <DcrDeratingEnabled>False</DcrDeratingEnabled>
      <VinRangeLimitsDisabled>False</VinRangeLimitsDisabled>
      <VoutMaxLimitsDisabled>False</VoutMaxLimitsDisabled>
      <IswMaxLimitsDisabled>False</IswMaxLimitsDisabled>
      <StandardResistorValuesDisabled>False</StandardResistorValuesDisabled>
      <FswRangeLimitsDisabled>False</FswRangeLimitsDisabled>
      <ConstantCurrentLoadEnabled>False</ConstantCurrentLoadEnabled>
      <LTspiceTranEngineEnabled>False</LTspiceTranEngineEnabled>
      <LTspiceLoopEngineEnabled>False</LTspiceLoopEngineEnabled>
      <NumOfSchottky>1</NumOfSchottky>
      <PartNameControlStage1>NTTFD2D8N03P1E_Q2</PartNameControlStage1>
      <NumFetsControlStage1>1</NumFetsControlStage1>
      <RdsonControlStage1>0,0024</RdsonControlStage1>
      <Rdson1ControlStage1>0,0024</Rdson1ControlStage1>
      <Vgs1ControlStage1>4,5</Vgs1ControlStage1>
      <Vgs2ControlStage1>10</Vgs2ControlStage1>
      <QgControlStage1>9,3E-09</QgControlStage1>
      <QgsControlStage1>3,7E-09</QgsControlStage1>
      <QgdControlStage1>1,6E-09</QgdControlStage1>
      <CossControlStage1>4,98E-10</CossControlStage1>
      <Rdson2ControlStage1>0,0018</Rdson2ControlStage1>
      <Qg1ControlStage1>9,3E-09</Qg1ControlStage1>
      <Qg2ControlStage1>2,05E-08</Qg2ControlStage1>
      <VmillerControlStage1>2,6</VmillerControlStage1>
      <VthControlStage1>3</VthControlStage1>
      <RgControlStage1>0,75</RgControlStage1>
      <VendorControlStage1>ONSEMI</VendorControlStage1>
      <VdssControlStage1>30</VdssControlStage1>
      <VdiodeControlStage1>0,8</VdiodeControlStage1>
      <LengthControlStage1>3,3</LengthControlStage1>
      <WidthControlStage1>3,3</WidthControlStage1>
      <HeightControlStage1>0,75</HeightControlStage1>
      <PackageControlStage1>WQFN-12</PackageControlStage1>
      <PartNameSyncStage1>BSC009NE2LS5</PartNameSyncStage1>
      <NumFetsSyncStage1>1</NumFetsSyncStage1>
      <RdsonSyncStage1>0,00095</RdsonSyncStage1>
      <Rdson1SyncStage1>0,00095</Rdson1SyncStage1>
      <Vgs1SyncStage1>4,5</Vgs1SyncStage1>
      <Vgs2SyncStage1>10</Vgs2SyncStage1>
      <QgSyncStage1>2E-08</QgSyncStage1>
      <RgSyncStage1>1</RgSyncStage1>
      <VdiodeSyncStage1>0,51</VdiodeSyncStage1>
      <VendorSyncStage1>INFINEON</VendorSyncStage1>
      <VdssSyncStage1>25</VdssSyncStage1>
      <QgdSyncStage1>4,9E-09</QgdSyncStage1>
      <CossSyncStage1>1,4E-09</CossSyncStage1>
      <Rdson2SyncStage1>0,00075</Rdson2SyncStage1>
      <Qg1SyncStage1>2E-08</Qg1SyncStage1>
      <Qg2SyncStage1>4E-08</Qg2SyncStage1>
      <QgsSyncStage1>6,7E-09</QgsSyncStage1>
      <VmillerSyncStage1>2,3</VmillerSyncStage1>
      <VthSyncStage1>1,6</VthSyncStage1>
      <LengthSyncStage1>6,35</LengthSyncStage1>
      <WidthSyncStage1>5,35</WidthSyncStage1>
      <HeightSyncStage1>1,1</HeightSyncStage1>
      <PackageSyncStage1>PG-TDSON-8</PackageSyncStage1>
      <PriceSyncStage1>1,415</PriceSyncStage1>
      <CostSyncStage1>1,415</CostSyncStage1>
      <PartPriceSyncStage1>1,415</PartPriceSyncStage1>
      <NumFetsControlStage2>1</NumFetsControlStage2>
      <NumFetsSyncStage2>1</NumFetsSyncStage2>
      <PriceSyncStage2>1,415</PriceSyncStage2>
      <CostSyncStage2>1,415</CostSyncStage2>
      <Efficiency>0,9887</Efficiency>
      <VoutRipplePercent>0,0100901741</VoutRipplePercent>
      <LossCurveA2>0,0032313224</LossCurveA2>
      <LossCurveA1>0,1083926332</LossCurveA1>
      <LossCurveA0>0,2920569269</LossCurveA0>
      <LossCurveR2>0,9932995994</LossCurveR2>
      <IlimMax>0,05</IlimMax>
      <InputLCFilterEnabled>False</InputLCFilterEnabled>
      <NumCf_in>1</NumCf_in>
      <NumCd_in>1</NumCd_in>
      <Rd1_in>-1000000</Rd1_in>
      <Rd2_in>1000000000</Rd2_in>
      <NumCd3>1</NumCd3>
      <Rd3>-1000000</Rd3>
      <PartPrice>2,37</PartPrice>
      <PartCost>2,37</PartCost>
      <InductorPrice>1,36</InductorPrice>
      <CeramicCapacitorPrice>0,4293</CeramicCapacitorPrice>
      <CinCeramicPrice>0,4526</CinCeramicPrice>
      <InductorCost>1,36</InductorCost>
      <CeramicCapacitorCost>2,1465</CeramicCapacitorCost>
      <CinCeramicCost>0,4526</CinCeramicCost>
      <PartFamily>Switching Regulator</PartFamily>
    </OutputRailData>
  </OutputRailDatas>
  <NotationLabelControls />
</Root>

Peltier-Drivers.zip

Thank you for your advice.

Best,

+1 Achim
on Jun 11, 2024 1:43 PM in reply to Tekemachi
Hello Tekemachi,

thanks for the detailed data. But I have neither LTPowerCad II nor Altium running on my system, so I cannot look in depth into it.

Regarding your further description:

Tekemachi said:
I have tried removing C144 and put 3 more C145 but still, the circuit rings at about 1A of load current.

But it is working without an overheated Cout? That's a first success at least :-)

Tekemachi said:
it seems stables but rings horribly when I connect the load (about 1A)

The ringing you describe and show in your new scopeshot could be caused by a "nervous" feedback loop. You may be able to compensate this by modifying the network connected to the ITH-pin.

The chapter "Checking Transient Response" in the datasheet may give you a hint for an optimization by measurements. And I think LTPowerCad II also allows you to create some Bode-plots of the feedback loop for various values of the compensation network. Maybe you can generate some additional phase reserve at the critical frequency of your feedback loop.

best regards

Achim
Cancel
Up +1 Down

Reply

Reject Answer

Cancel

Partner Forums

Partner Libraries

Products Mentioned

Output not stable